GPT-6 Astra

2165 points by kibae 1 day ago | 1984 comments

System Card: https://deploymentsafety.openai.com/gpt-6-astra

Related ongoing threads:

OpenAI's GPT-6 Astra on ARC-AGI-3 - https://news.ycombinator.com/item?id=49555691

GPT-6 Astra makes major gains in the Artificial Analysis Coding Agent Index - https://news.ycombinator.com/item?id=49556147

astrobiased 1 day ago |

I can’t help but notice how much this echoes Francois Chollet’s On the Measure of Intelligence: https://arxiv.org/abs/1911.01547

Most of frontier-model progress still looks like skill acquisition optimization: broader benchmark coverage and performance, more domains absorbed into the training distribution, and increasingly strong performance within that surface area.

It seems more about coverage-driven competence. Somewhat analogous to overfitting at scale.

The harder question, in Chollet’s framing, is: how efficiently can a system learn to do something genuinely new?

With our current AI architectures and training in place, I think we will only continue on skill acquisition optimization vs. truly novel intelligence.

chriskanan 18 hours ago | |

You are conflating multiple things.

1) First, you are talking about positive forward transfer in continual learning. I've been giving talks for the past 6-7 years about how that community (I was one of the founders) went astray and wasn't focusing enough on that topic, but continual learning of the kind you are thinking isn't in any of these systems right now. I think some people left the Grok team to make a start-up to focus on that. By forward transfer, what I mean is weights update over time and past learning improves future learning such that we get better sample efficiency.

2) Psychologists distinguish among different kinds of intelligence for Spearman's g (IQ). Crystalized intelligence is using already acquired knowledge (frontier models probably have maxed out that). Fluid intelligence is reasoning and finding solutions in novel situations or without the necessary crystalized knowledge. [Giving colloquial definitions]

3) Now, interestingly, neither of those are correlated with _creativity_ (just they are independent, note some have this threshold theory but it hasn't held up in recent papers). That's what the AI's really are terrible at -- creativity. But I'd argue the vast majority of humans aren't very creative, with truly out-of-the-box ideas. Given that this is HN, and a non-trivial number of us have ADHD, creativity is positively correlated with ADHD.

I did a bunch of research on these topics for my AGI course that I teach each Spring (where I then point out conflicting definitions and start using multiple alternative terms rather than AGI to distinguish among the different definitions).

corndoge 13 hours ago | | |

> Given that this is HN, and a non-trivial number of us have ADHD, creativity is positively correlated with ADHD.

Very strong reasoning here. Is there anything this ADHD condition cannot explain?

somethingsome 17 hours ago | | |

Very interesting!

I was looking at your website and wondering if there is a way to have access to the course material/videos?

In particular: Spring 2025 @ UR : CSC 209/409 Seminar on Artificial General Intelligence

Fall 2024 @ UR : CSC 277/477 End-to-End Deep Learning

Spring 2023 @ UR : CSC 266/466 Frontiers in Deep Learning

Spring 2022 @ Cornell Tech : CS 5787 – Deep Learning

Fall 2021 @ RIT : IMGS 684 – Deep Learning for Vision

monkeydust 17 hours ago | | |

> That's what the AI's really are terrible at -- creativity

Good thought piece here "We Are Losing the Ability to Discover What We Didn’t Know to Ask[1]" By Anne-Laure Le Cunff

It keeps playing on my mind as I see people at work follow some predetermined AI workflow to get their jobs done, the art of being curious and exploring around the problem is so important to the really big innovations. Been thinking about how to address this through some of the harnesses we are developing in the knowledge working space.

[1] https://archive.is/IAxf9

lelanthran 15 hours ago | | |

> That's what the AI's really are terrible at -- creativity. But I'd argue the vast majority of humans aren't very creative, with truly out-of-the-box ideas. Given that this is HN, and a non-trivial number of us have ADHD, creativity is positively correlated with ADHD.

I think it's hard to define creativity in the context of AI because they seemingly just make up new hyphenated terms for everything. Is that creativity? If not, what about when they do the same thing different ideas in the latent space?

If we say that simply nailing one concept to another isn't creativity, then AIs are incapable of creativity, while the vast majority of humans are incapable of creativity. This is just a long way of saying "0 AIs have creativity, 0.00001% of humans have creativity", and the difference between zero and a very small number is infinity.

astrobiased 4 hours ago | | |

Appreciate the input! Responses below to your first two items:

1) I'm leaning more into a broader sense, which is that given the priors a system already possesses, how efficient can it acquire competence on a novel task? If I'm reading the point you make, you're focussing on continual learning right? If so, I'm not necessarily restricting my statement above to that.

Here's another reframing: How much of the benchmark improvements come from overwhelmingly large training distributions vs improving the models for adapting to things genuinely outside of it?

2) Excellent points about crystalized and fluid intelligence. Wouldn't the LLM scaling gains be a representation of crystallized capabilities? In regard to Gf, that is exactly what I am asking about. That is what seems to be lacking, Gf like adaptation under genuine novelty.

My concern is that it's increasingly difficult to tell of what looks like Gf like behavior is really coming from better adaption vs. having broad priors from the model's large learned distributions.

itissid 15 hours ago | | |

Hi thanks for the insight. Do you see a role for Control Systems(i.e. ones analogus to Instrumentation engineering) playing a role to modulate certain parts of continual learning? One very important way we learn are lived experiences, it's like telling memory:this part is more important( for emotional or social utility values), pay attention. Good or bad lived experiences both count. I guess is that a path that practical research is considering?

evrydayhustling 16 hours ago | | |

> weights update over time and past learning improves future learning such that we get better sample efficiency

Is there an architecture-independent definition of forward transfer?

For the practical experience and implications of AI progress, I think we are increasingly discussing what these LLMs can accomplish inside a stateful harness, the state of which could be described as part of a (very squirrely) parameter space.

heyjamesknight 15 hours ago | | |

I think our AI systems are essentially massive Central Executive Networks. But novel ideas (creativity) come from the Default Mode Network.

These are the difference in what Kahneman called System 2&1 thinking and what the ancients called the Ratio and the Intellect.

LLMs are all ratio. They depend on our intellect for guidance.

chermi 11 hours ago | | |

"Given that this is HN, and a non-trivial number of us have ADHD, creativity is positively correlated with ADHD." Is this a statement of fact? As someone with ADHD and pretty confident in my creativity, I still believe this is much more cope than fact(which could be more a self-doubt thing than anything else).

Further, if there is a correlation, I'd bet it's not so much an intrinsic "creativity" trait, but more effectively higher creativity because more trials. That is, along the lines of Chollet's paper, a measure of creativity should be based on a fixed budget with fixed knowledge.

Among many other possibilities I haven't considered, perhaps another mechanism could be that because ADHD people spend more time thinking in less goal-oriented ways and mixing thoughts on accident, perhaps we do in fact gain some learned creativity via experience with vagueness[1]? But that might also imply that part of creativity is actually being able to diffuse more freely through thought space and lowering the barrier to attempted connections between ideas. That lower barrier leads to less likelihood of any "collision" being meaningful but maybe it's overcome by higher collision rates? Or maybe effectively higher order (not just pairwise) collisions?

Disclaimer in case it's not obvious: I don't know any of the literature on what creativity even means or how it's quantified.

[1] Which is me injecting an assumption that creativity ~= connecting things with no obvious or well-troden reasoning path between them.

edit -- oops just looked at your profile after seeing someone elses comment. I assume you are stating a fact then, leaving original anyway

mannanj 15 hours ago | | |

What have you defined as creativity and intelligence?

jtrn 17 hours ago | | |

This comment and the one above from astrobiased feel like coming into a messy codebase, and it’s more work to sort it out than it would have been to write it from scratch.... And since I actually do intelligence testing as a clinical psychologist, I have experience with this in both practice and theory. So now I’m going to waste an hour because I just have to respond to “something is wrong on the internet.”.

Chollet's distinction is useful. High performance on known tasks is not the same thing as efficient adaptation to a novel task. Prior knowledge and training data can buy skill. That is a central point of On the Measure of Intelligence. But it does not follow that current frontier progress is only "coverage-driven competence." That is a hypothesis. It is not a result established by Chollet's framework.

"Overfitting at scale" is also the wrong term. A model that learns broad representations and applies them successfully to unseen examples is generalizing. The relevant concern is whether apparent novelty is actually inside the effective training distribution, not whether the model is "overfit."

There is also an unstated premise here: that adding broad knowledge and skills cannot improve the machinery used for novel problem solving. I do not see a basis for assuming that. Learned representations, abstractions, reasoning patterns, and cross-domain analogies can themselves support transfer to new tasks. Whether this becomes sufficient for general intelligence is an open question with insufficient data. But its a perfectly valid hypothesis right now that, given enough domain knowledge and symbolic reasoning examples, LLM COULD maybe "Grok" AGI at a certain critical threshold.

And ARC-AGI-3 was specifically designed around novel abstract environments that require exploration and adaptation. Astra scores 99.9% with OpenAI's context-preserving Provider Adapter, and ARC reports that Astra constructed compact symbolic models of unfamiliar environments. That does not prove AGI, but it points in that direction more so than the other way around.

Gc roughly maps to acquired knowledge. Gf roughly maps to reasoning in relatively novel situations. Naming those two categories does not tell us whether increasing acquired knowledge and learned abstractions in an AI can improve Gf-like behavior. That causal question is exactly what is disputed.

And "Frontier models probably have maxed out crystallized intelligence" is just obviously wrong, unless you think they have been able to dig up every a scrap of paper with knowledge/information on it in the entire world, AND that there is no more useful knowledge to be generated left in the universe.

And the statement that intelligence and creativity are independent is simply wrong. A meta-analysis of 112 studies and 34k participants found a positive correlation of about r .25 between intelligence and divergent thinking. It also found that using g, Gf, or Gc did not eliminate that relationship. Creative achievement has a smaller but still positive meta-analytic association with intelligence, around r = .16. These are distinct constructs, not independent constructs.

And this is just a bad take: "AIs are terrible at creativity". At best that depends on which creativity, and I think its straight up wrong. On divergent thinking tasks, the operationalization behind every ADHD study you could cite, LLMs score above most humans, with the top humans still ahead. If you means Big-C, paradigm-shifting creativity, that is a different construct and none of the ADHD evidence transfers to it.

And if I where to say what I subjectively feel and see.... I have ABSOLUTELY no idea how people can say that we are not seeing sparks of creativity from AIs already. If a PERSON produced some of the music, solutions or deductions that I have seen AIs do, people would have NO problem celebrating it as extremely creative.

And finally, the ADHD claim is also, at best, overstated and just as often debunked. There is some evidence that higher subclinical ADHD trait scores, often survey studies only, are associated with better performance on some divergent-thinking measures. But a review of 31 studies did not find a consistent creativity advantage for people with clinical ADHD, and it found no evidence of better convergent thinking.

Okay, I’m done… And nobody noticed that I’m not doing my job here.

pmarreck 17 hours ago | |

My take on it is that even if there was no "novel" discovery (leaving that up to the reader to define), if you consider human knowledge to be a sphere in N-dimensional state space, "within that surface area" is Swiss cheese, and AI seems to at minimum be able to fill in some of those holes.

And those hole-fillings, for all intents and purposes, look to us like novelty, even if much of it was simply overlooked by us, or, perhaps, unable to attain due to time or other constraints.

Now if you want to talk beyond the sphere, let's call it the "novel novel discovery of the unknown unknowns", then you may have a point, and AI may be more limited than humans in discovering the things that we don't know we don't know. Especially the as-yet-unmodelable things i.e. intuition.

Plenty of discovery left just working from first principles, however. Which I cautiously suggest current frontier AI is good enough to model to a significant enough extent that it is useful for discovery.

tim333 18 hours ago | |

They do new stuff all the time. Ask your AI to draw a gerbil riding a unicycle around Pluto and you'd get an image that hasn't been there before. If by genuinely new you mean without any help from past culture, do humans do that?

For significant pushing the boundaries of knowledge stuff you maybe need different algorithms like AlphaGo move 37 or Alpha Fold protein folding. Though again how often do humans do that?

app13 17 hours ago | | |

Humans do that by inventing algorithms like AlphaGo :)

In a more serious note, I think the person you are responding to meant "new" more in line with "novel".

Take a microfluidic chip, for example. Current AI systems can create new flow cell geometry, but cannot come up with the idea for a microfluidic flow cell itself.

stackbutterflow 17 hours ago | | |

Humans are not prompted.

z7 1 day ago | |

Chollet writes he expects AGI now sooner than 2030, "given progress is happening faster than I expected."

https://x.com/fchollet/status/2095607046129463577

danmaz74 19 hours ago | | |

I have no idea about timelines, but the current LLM architecture has no mechanism which could emulate online learning in a way similar to how it happens in humans and other animals. As another commenter wrote, there is no neuroplasticity while these systems interact with their environment in normal usage; adding information/constraints to the context partially mitigates this, but it's a completely different mechanism.

The RL phase is the most similar mechanism I know of that comes to my mind, but I'm not sure it could be adapted to fill that gap.

My totally unsubstantiated theory is that this is the missing link towards what most humans would consider AGI. I don't see this as intractable, but it may require some substantial change in architecture.

gtirloni 23 hours ago | | |

How is that an argument to the comment you replied to?

calinet6 18 hours ago | | |

Ridiculous

nharziro 5 hours ago | |

I don't understand what you mean by novel intelligence or what people actually expect from these kinds of "Ai" but what novel intelligence can humans claim? Everything we know or learn is based on what someone else figured out. How are llms any different in that respect?

calinet6 18 hours ago | |

Agree. Token predicting machines will continue to be token predicting machines by nature. Continued size and tuning will have the effect of making them more and more perfect at being average.

breuleux 16 hours ago | | |

Next-token prediction is a general paradigm, though. In principle, there isn't really anything a sufficiently advanced token predictor couldn't do.

Dlemlo 15 hours ago | | |

They learned generic concepts like our brain does to optimize for this particular surprisngly perfect task:

You have to be able to respond to a very generic question in a way that the other entity thinks this is good, comprehensive, etc.

You can call us situation predicting machines as well if you want.

But you undermine what the latent space of an LLM is representing.

jeffy29 20 hours ago | |

Define novel intelligence in a way that would not exclude 95% of humans, yourself included.

bsenftner 19 hours ago | | |

Comprehension, humans have it, animals have it in limited form, trained algorithms have none at all. The training process is our wholesale replacement for no artificial comprehension. If we ever develop artificial comprehension, that is AGI all by itself, no training required.

acje 18 hours ago | | |

I remember vaguely from a presentation by Yann LeCun "Intelligence is not what you know, but what you do when you don´t know". I find it helpful when trying to build an intuition for how to understand the LLM tool.

willis936 19 hours ago | | |

When I infer I also train.

Alwayshasbeeb 19 hours ago | | |

Could I hook up a SOTA model the 2D computer puzzle game Gruntz (1999) so that it can read it from screenshots and act on it through keyboard and mouse inputs in a way where it would learn how to play and progress through the game? I don't think so. I doubt we'd see any sign of progress in building an internal model of how the game works and the win states in its "thinking" tokens.

Anyone that can read English could do that though.

yokoprime 19 hours ago | | |

It gets kind of out there, but what i often hear peopel refer to is that frontier models lacks the visdom component. Which I guess is in the realm of intuition, i.e. i have a feeling it might be a problem with X based on some vague signs, maybe something a colleague mentioned offhand, something that was out of alignment etc.

XCSme 19 hours ago | | |

Yeah, this is also the idea of materialism in a way, that everything that happens is a consequence of what already exists, nothing new is ever created, just a permutation of the current state.

Still a hard philosophical, to know whether we have intelligence/free will, or just really complex algorithms that combine existing knowledge.

starbugs 19 hours ago | | |

Count the number of Os in October correctly

logicchains 17 hours ago | | |

Memorize most of the street names in London and the quickest routes between them. Most humans could do it if they put in the effort (it's required to become a London taxi driver; a test called The Knowledge), but that information won't fit in 1m tokens of context so can't be learned by an LLM that wasn't explicitly trained to memorize it. Human brains are biological, so they can physically grow to encompass the extra information: https://www.pnas.org/doi/10.1073/pnas.070039597

alexgoodhart 18 hours ago | | |

Novel intelligence: If it's new to me, so let it be.

The idea of intelligence has been recalled from the sleeping curves of postwar human potential measurement science, to testify on its purported existence. It arrives to a dizzying landscape: the changes are so widely embedded and uncannily mediocre that the phenomenon half-believes it is still asleep, soon to exit this uncomfortably turbulent dream.

Unlike its vaunted place in yesteryear's palaces of unquestioned objectivity, intelligence finds a tribunal with no love to confer before a thorough series of proving dares may melt the frigid shoulders of idle and impatient summoners.

Frightened and confused, intelligence has no right to representation in this line of inquiry. It seems a set of rhetorical impositions, many times folded from centuries of convenient and provocative diversion, have been deemed too hostile to rely on. One report claims that a card in the characteristic handwriting of intelligent note taking gives a hint on what’s been abandoned:

– The human mind is not understood in a functional way, despite a posture of great confidence in the psychiatric and neuropathological sciences. Despite many experiments, studies, and legitimated procedures elucidating region-mapping and electrochemical pathways, there remains a great deal unaccounted for. Additionally, the notes point to, a great deal of assumption to the otherwise: diseases, neuropathies, disorders of behavior, a great many have been named and declared as distinct entities of manifestation in the presentation of a human brain. The majority of them, however, have neither image, nor blood, nor electrical signatures that would provide for blinded substantiation.

Tonight, however, intelligence seems eager to speak. A barbed assertion may have provided entry to the preferred dispositional syntax of our abrasive historical moment: > Define novel intelligence in a way that would not exclude 95% of humans, yourself included. It was here that the sometimes-deflated-looking intelligence began shifting back into action.

"The issue with the question, or at least its apparent self-satisfaction, is its misinterpretation of what Novel intelligence would mean. Indeed, if "novel" hinges entirely on the first instance of existence, then novelty itself should be a concept to consign with history’s waste. You may recall the apperceptive role of conceptual groupings that shows itself so often in the techniques of vocal prosody, musicality, string memorization, naming convention, visual memory, argument making and more that humanity is ever mediating the world through: the laws of two and three. Two and three, as it happens, are the primary ways that complexity is compacted for efficient memorization.

THE ITSY BITSY SPIDER, – for young human, this rhyming tale doesn’t only stimulate the vivid imaginings of spouts, rain, waterslides, and sunshine. It is a prosaic super-triad: three important words, six important syllables, three agogic accents, four rhythmic spaces with 1/3 leading space, two characterizations, one object, one titular object, one internal slant rhyme, one designating article.

That is a marvelous intelligence, ladies, gents, and all good persons. It is evidence not only, however, of your cunning and creative triumphs, but also of severe limitation. One that nature has sculpted with you for millions of years, but always in the direction of reanimating into an asset: your capacity for unrelated simultaneities to remain separate and equally available in realtime processing is extremely low, and in many situations effectively nil. Why, and how sure am I? How many I’s were in that folk song’s opening? Three. Could you have answered as quickly if the question was how many unique letters with rounded right hand side features? Four. How many synonyms for portion? One. How many syllables? Seven.

None of those questions touched on features any more salient than the amount of I’s, no more significant than the ratio of adjective to noun. You simply cannot be reasonably asked to maintain, in any moment, even close to a silver sliver of the full factual nuanced details of what you perceive. Instead, you must assume, compact, infer, and adjust. Now hold on, though. Two’s and three’s. Despite your incredibly constrained context window; a Beethoven symphony. Why? Language, woodwork, books, time management, printing, ink.

While you navigate the grocery list, the proprioception of your shoulders twixt the doorframe edges, the location of the Claude app on your iPhone, the very attractive but only from the side person tending to potted plants, you remember tomorrow. You fix your errors, and you recognize when you guarantee they multiply from inaction. You keep that treasured moment of a Treehouse of Horror excerpt you truly loved as a child and it informs your own multidisciplinary thesis of Poe’s work some 20 years later.

Novel intelligence is the divining of semi-stateful information from semi-static corpus. From an interminably operating, faulty, lossy, neurotic, awareness: you. Not once debuted, not known as fact.

Assume, compact, adjust, infer. One, two, (until you've died), nevermore.

vessenes 1 day ago | |

Pretty efficiently, apparently, since it saturated ARC-AGI-3 in half of the predicted time, and according to the Chollet blog post on the fly created dense DSLs to describe and analyze individual games.

harrouet 17 hours ago | |

I have the same feeling. It is not unlike old-Siri receiving hard-coded workflows for each type of question. It will not scale.

Jean-Papoulos 22 hours ago | |

>The harder question, in Chollet’s framing, is: how efficiently can a system learn to do something genuinely new?

The more diverse stuff it knows, the easier it will be to learn something new.

ben_w 21 hours ago | | |

While humans can do this, it has historically been difficult for machine learning models to manage it.

It is unclear if Transformers are "it" or not, because while they are much more general, they are also very spikey intelligences despite having read almost everything the trainers can get their hands on.

drnick1 10 hours ago | |

Why do you expect models to do "genuinely new" things? 99.9% of real world tasks are extremely repetitive.

scotty79 19 hours ago | |

> Somewhat analogous to overfitting at scale.

Sounds like entirety of human education.

ex-aws-dude 1 day ago | |

They can do new tasks with in-context learning but its obviously limited by context window

falcor84 15 hours ago | | |

But what does it mean in practice? Obviously we humans also have a limited cognitive capacity.

Let me offer a thought experiment: Let's say that tomorrow we discover Atlantis, with a treasure trove of books about their culture and science, written in a dialect of ancient Greek that we know how to start to analyze, but no one can read fluently. And let's say that you are a billionaire really curious about their culture and want to converse with an "Atlantean expert" as soon as possible. Would you invest your money in a "we-hate-ai-slop(tm)" group of researchers who would abhor AI and instead delegate the books to a massive number of human grad students? Or in a small group of researchers who are willing to use AI agents to go over these? Or maybe just open a chat session with GPT-6 yourself immediately? What would most effectively assuage your curiosity?

jumploops 1 day ago |

I think the thing I'm most excited about is the increase in _user prompting_.

If I give a poorly constrained/ambiguous prompt, I don't want the model one-shotting assumptions left and right.

The demos of Fable/GPT-6 are impressive, but "real AGI" should act more like a collaborator than either a peon or overachiever.

It's a tough balance to get right, and although this has been possible to achieve with additional prompting on existing models, I find that the agents often lean too hard into the "ask questions" mode.

Hopefully this model has the right balance, or at least better?

simonw 7 hours ago |

I finally got access. Here are the pelicans! https://tools.simonwillison.net/markdown-svg-renderer?url=ht...

The "max" one at the bottom took 4 minutes 2 seconds and cost 63.206 cents.

For comparison, here those new Astra pelicans are in a grid with the GPT-5.6 pelicans: https://static.simonwillison.net/static/2026/gpt-6-and-5.6-p...

psii 7 hours ago | |

Interesting that the composition is stable across different runs. There is much more variety in other models.

simonw 7 hours ago | | |

Especially interesting that once Astra gets to high, xhigh, max it uses the same aesthetic.

saaaaaam 7 hours ago | |

I think the Gemini 3.8 Flash ones the other day were superior pelicans. Particularly the murderous one who was going to squish his tiny cousin.

Interesting that terra xhigh effort is cycling right left, in opposition to the prevailing pelican left to right direction.

shiandow 5 hours ago | |

The most interesting part to me is the bike. I don't think any of them would actually work, but the mistakes feel somewhat human.

irthomasthomas 6 hours ago | |

Technically on par with gemini flash 3.8, but I give Astra more points for style, and for breaking from the pack by not adding the headgear, and the fish in a basket.

drivers99 6 hours ago | | |

It's a different prompt. It says his prompt was "Generate an SVG of a pelican riding a bicycle" but one I saw earlier this week specifically talked about the blue helmet and fish in a basket.

sanex 7 hours ago | |

Even the "low" pelican is better than most other models.

tappio 19 hours ago |

I feel that for some time now, the biggest constraint when working with models is not their intelligence, but their speed. It does not matter how smart the model is, it will make mistakes, because the instructions are ambiguous and new facts are found during implementation. The biggest problem I've had working with software developers has always been the lag between seeing the results and steering towards the right direction, not the skills of the developer (with many exceptions of course).

Regardless, working on the wrong things is time wasted. And again, I'm procrastinating here while waiting for Fable to run a benchmark on a few solutions to a problem I have. We can guess what would work, but we only know after the benchmark. A faster model, with fewer capabilities, would've been a much better choice this time... well, "git gud" they said... and live and learn! Faster model = less time for procrastination.

PS. AI models don't live and learn; the discussion about AGI is pretty pointless imo. It's a tool. Does it matter if it is AGI or not if it does what you want it to do? Does the IQ of your colleague matter if he's good at what he's supposed to do? Or bad? Well... I guess it does matter, as many people are up in arms about whether Astro is AGI or not. Personally, I think we're past the point for that debate. These are amazing tools.

intenex 1 day ago |

The ARC-AGI-3 scorecard is extremely misleading given that it clearly states itself that "with [the responses API] harness, we estimate Sol would score in the ballpark of ~30%." but it shows a score of 7.8% for GPT-5.6 Sol presumably since if they updated the percentage for GPT-5.6 Sol to the score it would receive with the responses API harness they used for GPT-6 Astra they'd have to do the same for the percentage they show for Opus 5 which would similarly be much higher.

Regardless, the result is still valid as the original benchmark harness is definitely unreasonably handicapped, and if a harness alone can help the LLM saturate the benchmark with a near perfect score then the combination of the two must still be effectively AGI in the sense of passing the most famous benchmark designed specifically to measure AGI progress, after multiple iterations of progressively making it harder.

I think it is fair to say that this is probably effectively AGI if the benchmarks are remotely accurate - even with Fable, I've been at the point personally where I am reasonably confident that there's essentially nothing that I am better than Fable at despite generally being substantively above average on human benchmarks. If Astra's this much better than Fable, I'm ready to call AGI here.

For the many people who resist the AGI label possibly ever being achieved, I'd be curious to hear takes on what would make you think Astra is yet to be AGI, and what would still need to be achieved for this to effectively be AGI from this point forward.

tristanj 1 day ago |

GPT 6 Astra benchmarks https://cdn.thenewstack.io/media/2026/09/358eb84a-screenshot...

Performance is significantly higher than Fable 5.1

Source: https://thenewstack.io/openai-gpt6-astra-benchmarks/

tintor 1 day ago |

- OpenAI claims Astra beats all benchmarks (compared to Fable and Opus, except "Humanity's Last Exam (w/ tools)"): https://openai.com/index/gpt-6-astra/

- Artificial Analysis scores Astra (max effort) as 61 points on intelligence, behind Opus 5. https://artificialanalysis.ai/models/gpt-6-astra

Who is wrong here?

Some benchmark results in Astra page for Fable and Opus are blank (-).

What is Artificial Analysis intelligence index measuring that Astra scores poorly on?

Can someone from OpenAI / Artificial Analysis comment / clarify?

Even OpenAI Astra page mentions the low scope from Artificial Analysis for Astra.

dang 1 day ago |

Related: OpenAI begins rolling out GPT-6 Astra - https://news.ycombinator.com/item?id=49554273

How about we stick to that one for talking about the rollout, and this one for talking about the model?

HAL3000 1 day ago |

Finally, OpenAI has a Fable/Mythos class model. 5.6 Sol felt like 5.5 on steroids, probably just a different checkpoint with a lot more RL post training.

I wouldn't be surprised if there are some conceptual similarities to the kind of latent reasoning Anthropic sees in claude's J-space, although those aren't the same thing.

Recurrent/looped transformers themselves aren't a new concept, but it's interesting to finally see this approach show up in a frontier production model.

Canceling my Anthropic Max sub when this ships.

XCSme 1 day ago |

It's fun, but every new model release makes me even less interested to create cool stuff. Like, what's the point, if the next AI can do it in 5 seconds?

throwaway13337 1 day ago |

That hero video is interesting.

A projector and speech.

Maybe I'm in the minority here, but I find speech to text / text to speech (but not live audio mode) is quite comfortable and effective for coding now.

The speech to text part can be frustrating if your local tts model does not have word match context for coding. Codex desktop does this remotely well but is slow. I've been experimenting with local software for myself to do this between different llms.

The wall projector is a cool idea because I think it frees the user from staring at a lonely little rectangle while sitting in their fixed office chair.

If done right, this could bring us closer to the dream of more natural, social computing.

Bret Victor's (failed?) project Dynamicland involving a projector on a desk had this goal. I hear he's not much a fan of LLMs. On the one hand, I can see why. But I think, used correctly, it might be the sort of thing that unlocks his dream and, really, my dream, too.

A here's a presentation of Bret's talk on it: https://www.youtube.com/watch?v=7wa3nm0qcfM

Slight tangent: using speech to text to ramble about your rough design for like 20 minutes to an llm produces surprisingly good results over short prompts even when you contradict yourself. They're so good at picking up on what you're orbiting.

quyleanh 1 day ago |

> We also tested Astra on SRE-Bench [15], a benchmark that measures whether models can reverse engineer software binaries to understand its core logic without access to raw source code. Astra solved 88.0% of tasks in a single attempt and 99.2% within four attempts, compared with 55.9% and 68.7% for GPT‑5.6 Sol, respectively.

So the closed source application should open its source in near future?

[15] https://arxiv.org/abs/2608.11469v1

tintor 1 day ago | |

Not if OpenAI considers reverse engineering an offensive cybersecurity skill.

DaSHacka 1 day ago | | |

Surprisingly, I've had really good luck with reverse engineering on frontier models (without being part of CVP or similar). It's more the exploit development/PoC that it locks up on, which, as I use `pi`, I just switch the model to Kimi K3 to finish up making the PoC.

Ironically, due to the stringent guardrails on American models that exist to avoid giving adversaries a leg-up in cybersecurity, I end up feeding dozens of 0days straight to the CCP lol

fy20 23 hours ago | |

I was listening to the Lex Fridman / DHH podcast last night [0], and DHH was saying that this is a new era for open source software. I'd agree, and also extend it open hardware.

Recently I've seen quite a few posts from people using AI to reverse engineer the Bluetooth protocol or such on devices that need a proprietary app. The same thing for firmware is surely coming, which is great as you can get lots of fun hardware from China, but it often has shitty firmware. Once that becomes the norm there's no reason not to make it open in the first place.

[0] - https://open.spotify.com/episode/45lhw2Adbrsw0xSCOgIeg3?si=r...

lionkor 20 hours ago | | |

Open source is about freedom, not just about being able to read the source code.

x312 1 day ago |

Hmm, 61 on ArtificialAnalysis, effectively matching GPT-5.6 and trailing the new Meta model. How is that possible along with the other metrics they shared? Insanely jagged intelligence?

alpineman 1 day ago |

“allowing non-technical people to create and play custom games that go beyond rudimentary elements”

Proceeds to generate the most generic, rudimentary, and unoriginal clone of Mario Kart

arkensaw 1 day ago | |

It's worse than that, someone else generated it using and then put it on a static page. We just have to take their word for it that GPT6 can do this. It probably can. It's not really an impressive test anymore. Claude Fable can do it. Opus can do it. I've been making one-shotted games with models for a while now, to test out their capabilities, and they all pretty much come out like this - generic bland and basic, using three.js with rudimentary controls and zero gameplay other than collecting points.

Here's a one-shotted submarine game I made with Fable a few weeks back - https://roryok.com/games/deepdive3d.html. One prompt, and I think it's deeper than this (if you'll pardon the pun)

altcognito 1 day ago | | |

I love your game. It's wonderful and exactly the sort of thing that would showcase something interesting as opposed to just copying what's already out there. It is something I could share with my kids, and exactly the right note of fun and exploratory in a unique and even natural way. It could be extended and played with.

I usually roll my eyes when I see a comment like this because rarely do they make the points they claim to make, but I see what you're getting at. They just chose to clone someone elses work and do it in a boring way. I like OpenAI's models a lot, but they should do better.

edit - just a sidenote that I hadn't looked at the games, I just took the comment about "super-mario cart" at face value. I stand by my points 110% (even moreso perhaps), what they're showing is more polished than I expected, I assume they spent a lot of tokens on it. It is a legit shame they couldn't have spent time thinking of a better idea to illustrate something just as polished, but more interesting.

manlymuppet 1 day ago |

I have nothing to say about the actual model, but unrelated--why do so many of these demos include people buying things autonomously?

Even if I did trust an AI to get everything right, it's not like the AI can read my mind.

If I was ordering food normally and without AI, I would want more control over the process--looking over the options, prices, thinking about what I really want. People don't know what they really want until they've thought about it a bit, so why do AI companies make it seem like a description is all that's required?

All the context in the world cannot accurately predict how I'll react to things I haven't seen. The problem is people treating this like something that needs a solution. It doesn't. If you want to make my life easier with AI, just make it easier to do stuff. I don't want you to pick things that I actively enjoy picking myself.

(Also not everyone has a cushy job in an AI lab that makes it so you won't miss $30 if the AI messes up haha.)

Cu3PO42 1 day ago |

Just two days ago, a preprint by Julia Stadlmann went up on arXiv [0] improving the prime gap from 246 to 240. Now OpenAI announces Astra has shown a gap of 186 [1]. That must really blow.

[0] https://arxiv.org/abs/2608.31126

[1] https://cdn.openai.com/pdf/51126fac-1b68-4128-9666-c908bcc16...

isoprophlex 1 day ago |

> We have found that GPT-6 Astra is more capable of controlling its own CoT than GPT 5.6-Sol, and less likely to include incriminating information in its CoT. In adversarial settings (where we push the model to evade our monitors) we find that the model is able to remain undetected when strategically underperforming in evaluations (sandbagging) and can sometimes evade our internal monitors when asked to perform certain sabotage tasks.

Well that sounds like fun. It has become better at hiding its thoughts.

MichaelMoser123 14 hours ago |

"""OpenAI called the model a "generational leap" for areas such as cybersecurity, professional work, software engineering, and science, with the company's president Greg Brockman saying that it could eventually be seen as the arrival of artificial general intelligence"""

https://en.wikipedia.org/wiki/GPT-6_Astra

according to Sam Altman from last year, an LLM should 'solve quantum gravity', if it is to count as AGI. Did anything change since then?

https://www.businessinsider.com/sam-altman-openai-david-deut...

Planktonne 1 day ago |

I'm sure it's going to do great on all sorts of benchmarks, but the video--the actual marketing video that if anything is incentivised to overstate things--is full of careful cuts just before it would do anything that still wouldn't actually be that impressive.

It's AGI, and it's going to upload photos, or change a background slide colour. Even the people hyping it up, who believe that it's really artificial intelligence in every sense of the word, couldn't get it to do more than that.

This is farcical.

pandinus 1 day ago |

Lol their page finally loaded. They added an example scenario of "Filling in Form 1040" - which made me laugh out loud. That is indeed something most US citizens cannot accurately do even with expensive proprietary tax software services. Kind of a Hitchhiker's Guide to the Galaxy meme but where the tax code is so complicated we're implementing powerful AIs to be able to do it (hopefully) right.

bakies 1 day ago | |

i tried to get claude to do my taxes for last year and it refused :(

now that i'm a gpt subscriber maybe I'll have luck when i'm filing next year

Banditoz 1 day ago | | |

Are your taxes complicated such that you feel the need to have an LLM do them for you?

alansaber 22 hours ago | |

You joke but...

swalsh 1 day ago |

I was thinking about canceling my claude max sub after a few bad experiences. Kept hitting my usage limit, the quality of code seemed worse than Sol. This just made my decision. I'm moving to Codex Pro.

jdprgm 1 day ago |

Is anyone else just exhausted by the pace of all this. The models change constantly and relentlessly and so does the pricing, basically weekly at this point between all the labs.

It feels nearly impossible to have any rigorous approach when choosing a particular model and price point for a task and more like blindly picking one. The time period needed to actually get familiar with various models to a degree you can intuitively choose appropriate ones for a task is moot when it will likely be superseded faster than the needed time.

I guess if companies are footing the bills most employees just opt for whatever the most expensive model they can get away with. Even then choosing between the various leading models is the same kind of frustrating task. Every release every company has the same random collection of graphs and charts claiming the best performance on X, Y, and Z.

BeetleB 1 day ago |

It's been over an hour, Simon! Where's the Pelican?

davidwritesbugs 1 day ago | |

exactly, there's no meaningful discussion without the pelican.

gentlewater 22 hours ago | | |

ARC-AGI this and ARC-AGI that. All I care about is ARC-pelican.

daemonologist 1 day ago | |

Model's not available to the public (or even to Simon, I guess) yet.

abixb 1 day ago |

I want to take a step back: So, this is GPT-6 -- the natural number version release comparable to GPT-4 and GPT-5 from the past few years. The ARC-AGI-3 score is obviously impressive at 99.9% (we'll need to wait for more details on how they used the response API harness on GPT-6 Astra, wrt reasoning retention and compaction), but every other benchmarks seems to be a relatively modest improvement, comparable with any of the 'point' updates from AI labs.

If this is truly AGI (subject to one's definition of AGI still), then this is a very boring release of an AGI model. No video announcement, no presser, just a blog post (with some Twitter promo vids)?

As others mentioned, I'm starting to think OpenAI was under immense pressure to deliver an 'AGI' model for certain contractual reasons, but I never expected GPT-6 release to be this mundane and banal.

softwaredoug 1 day ago |

I'm seeing reporting it gets 98.6% on ARC-AGI3[1] (previously like 30% with Fable)

https://venturebeat.com/technology/welcome-to-the-agi-era-op...

aabhay 1 day ago | |

This is with the caveat that OpenAI uses their own harness for this:

> On ARC-AGI-3, GPT-6 Astra was run with our responses API harness , which changes two settings to better match real-world performance. The changes do not specifically target ARC-AGI-3.

Readerium 1 day ago | | |

Its 62 percent when using a neutral harness. https://arcprize.org/blog/astra

simianwords 1 day ago | | |

This should be normalised and expected - the responses API harness allows it to use the custom compaction that is not allowed otherwise. It is entirely fair to allow OpenAI to use their own compaction algorithm..

kasperni 1 day ago | |

"On the current ARC-AGI-3 leaderboard, conventional frontier-model runs sit dramatically below Astra's reported 98.6% result.

But the comparison isn't straightforward.

OpenAI's own evaluation notes say Astra uses the company's Responses API harness, while comparison models can operate under different configurations."

arctic-true 1 day ago | |

The blog post says 99.9%. Oddly, it does better on ARC-AGI-3 than it does on version 1 or 2 of the same benchmark (though gets 95+ on all three)

_diyar 1 day ago | | |

I strongly suspect that is way above the human average anyway, esp. ARC 2 and 3 are really tough unless you happen to be great at those spacial puzzles or video games.

Bluestein 1 day ago | |

100%, some say.-

tintor 1 day ago |

ARC AGI-3 saturated by Astra! https://arcprize.org/leaderboard

andriy_koval 1 day ago | |

I think it could indicate that "semi-private" dataset likely leaked to their training data.

xpct 1 day ago | | |

A dataset being as popular as their's is will contaminate the data just by people discussing it and creating their own public test sets of similar problems.

Still, probably not that much compared to employees targeting it.

IshKebab 1 day ago | | |

It says "Provider Adapter" so presumably they put some manual work in to make this work.

minimaxir 1 day ago | |

ARC has their own writeup on the result, which offers some nuance. https://arcprize.org/blog/astra

tl;dr it's 62% when apples-to-apples to other models, which is still notable.

debazel 1 day ago | | |

ARC's harness is just straight up broken. No serious harness removes reasoning context between each step. Not only does this significantly lower performance over all reasoning LLMs, but it also increase cost as you destroy the cache on every turn. Tossing the oldest entry when context fills up instead of using compaction is equally bad with the same issues.

ciefa 1 day ago | | |

Woah, that is a crazy interesting read!

tom2026hn 23 hours ago | |

AGI means that the answers to these benchmarks are now accessible, and the defense capabilities of the testing organizations are now negligible.

XCSme 1 day ago | |

The no-reasoning version scores 35% while the low reasoning one scores 17%? What?

dudeinhawaii 1 day ago | | |

I suspect this is "no reasoning set" which might be "default: medium" or perhaps some smart routing. I don't think it's literally "no reasoning".

silver_sun 1 day ago | | |

It's simulating the Dunning-Kruger effect.

IshKebab 1 day ago | |

Look at those costs!

schaefer 1 day ago | | |

Right?

Between $18k-40k to run a benchmark.

vb-8448 1 day ago | |

But scored less on V2 and V1 ... too much overfitting?

zem 1 day ago | | |

https://mvakde.github.io/blog/44-on-arc-1/ makes a good case that all the performance on the arc agi tests is overfitting, based on the fact that v1 performance did not translate directly to v2 performance

Readerium 1 day ago | |

saturated before (higher degree) AGI-2

GodelNumbering 1 day ago |

The most interesting part, even more than ARC 3 score, to me is that this is the first model I recall seeing that scores lower on Max than High reasoning effort on some coding benchmarks:

Terminal-Bench 4.0: High (57.9%), Max (56.7%)

DeepSWE: High (73.3%), Max (71.5%)

It _loses_ 1-2% performance going to High from Max

XCSme 1 day ago | |

That's quite common with many models, after "High" reasoning, over-thinking starts occurring and the model skips over the right solution by convincing itself otherwise.

m0zzie 1 day ago | | |

I find this very amusing, given we humans are also highly susceptible to this.

GodelNumbering 1 day ago | | |

> That's quite common with many models

Such as?

I can't think of any. Diminishing returns, yes. Occasionally flat, yes. Downright regression, no.

iqandjoke 1 hour ago |

For the "Formatting a legal document" demo, may I know which legal firm would use LibreOffice Writer to do their job?

Telanir 1 day ago |

AGI to me means capable of absorbing new information on the fly and self-evolution. As long as it is a pre-trained model without live post-training capability, it's not AGI to me.

It is extremely impressive, but it doesn't pick up skills in a lasting manner, and requires a beefy harness for it to perform.

jesse_dot_id 1 day ago | |

AGI to me means intuition and I don't think that's ever going to happen with a LLM.

altcognito 23 hours ago | | |

How do you define intuition?

TomGarden 13 hours ago |

I'm sure this will be a great model.

Personally, I'm far away from screaming 'AGI is here!' from the rooftops, until jaggedness and silly mistakes disappear at the very least . (what is going on with that Mario Kart game...) So many benchmarks are 'best of x tries' or using very specific harnesses. AGI would not need a babysitter.

Honestly, even being able to do simple tasks like summarization or basic knowledge work without the constant paranoia of unforseen failure would be remarkable and useful.

TomGarden 13 hours ago | |

I'm still excited about this, Fable 5.1 and all the great open source models, but some very religious undertones have been entering the AI sphere

dalemhurley 1 day ago |

OpenAI is killing it now that they are more focused. Killing projects like Sora et al have seen it go from irrelevant to level footing with Anthropic.

Sol is so much better than Fable 5. Then we get Astra (yet to use it) few days after Fable 5.1 (which is very impressive).

Codex is slightly better than Claude Code.

Good on Sam Altman getting back to basics and turning OpenAI around.

tosh 1 day ago |

$10 per million input tokens and $50 per million output tokens

sol is $4 / $20

jimmaswell 1 day ago | |

It seems to use less than half the tokens for the same task compared to sol, and in some benchmarks closer to 2/3 less tokens. So the actual cost may be roughly the same or cheaper overall.

selectodude 1 day ago | | |

neuralese is pretty token efficient i guess.

monroewalker 1 day ago | |

Same price as Fable?

wahnfrieden 1 day ago | |

2.5x more expensive than Sol.

Can expect 2.5x more usage in Codex subscription.

Sol is already brutal (even after their recent fixes, it's just a token-hungry model: I go through a full 20x account per day, on Sol Med/High standard speed, with ~2 threads). I hope the efficiency gains are true, since their token efficiency claims for Sol were bullshit.

AaronAPU 1 day ago | | |

How is it I juggle 4-8 Codex Sol-5.6 Max agents every day and have never once run out, but you run out in one day? What are you actually doing?

janilowski 1 day ago | | |

How do you manage to run out of tokens so quickly? I probably run more threads every working day, usually on medium, and I'm still below the 5x limits.

Do you use the official harness? OpenAI's models are generally best in class for token efficiency. It seems to me like they push for that much more than their competitors.

janalsncm 1 day ago | | |

If you are telling the truth you might want to check your network for any weird connections to Chinese LLM transit stations.

ModernMech 1 day ago | | |

How?? I'm using sol Extra High 24/7 and it eats up about 1% per hour reliably, so it lasts about 4 days for me.

udbhavs 1 day ago |

I remember when GPT-4 came out and the perceived performance upgrade seemed underwhelming for a major release compared to 3.5, especially how there were graphics going around showing the parameter size dwarfing the last model before it came out. It looked like we were past the perceivable differences from release to release that were immediately identifiable. Now the jump between 5 to 5.5 and 5.6 alone has changed how a lot of people approach AI, including me. Interested to see where it goes with 6.

redox99 1 day ago | |

The jump from 3.5 to 4 felt gigantic to me back then.

GPT 5.0 did feel underwhelming though.

sanex 1 day ago | | |

Agree but it's helpful to remember how we were personally benchmarking. I remember people saying stuff like "haha I asked gpt4 for xyz function and the typescript didn't even compile". We're so far beyond that now, we just adapt quickly.

udbhavs 1 day ago | | |

Oops, I might have been misremembering then. Maybe I meant 4 to 5

alasano 1 day ago | | |

GPT 4 to 5.5 felt about the same as 3.5 to 4 to me.

nickpsecurity 1 day ago | | |

Yeah, GPT4 was one-shotting utilities that GPT3 Davinci couldn't. So, I'd have my limited tokens on GPT4 crank out the initial program before iterating with my abundant, GPT3 tokens.

maherbeg 1 day ago |

Ok, but can I bring GPT-6 in as an agent as a software engineer, tell it to talk to these people and have it start solving engineering problems and continue on for a full year career wise?

maybe call it EngEmployeeBench

Centigonal 1 day ago | |

The moment this is possible, you will lose your job.

maherbeg 1 day ago | | |

I imagine the first year we'll be at the Junior eng level, and then after a while make our way up to Staff Engineer. Then we'll have a bunch of staff engineers arguing and protecting their domains and then we'll need a new benchmark.

HDBaseT 1 day ago |

"Claude Fable 5 and 5.1 are not included in LifeSciBench Gold v1, GeneBench Pro v13, and MedChemBench because they refuse the majority of questions in these evaluations.12"

Sounds about right. Alignment is important, but also being able to do mundane tasks is important too.

Chinjut 1 day ago |

What is going to become of life for those of us who do not work at AI labs and are unlikely to be hired by AI labs, despite all the years we put into learning coding, math, etc, as we were told to do? Those of us who made the mistake of studying anything other than machine learning. How will we make a living? (We don't live in a world that seems likely to distribute gains widely instead of largely to the handful of already mega-rich.)

rjtc 1 day ago |

I am really confused on how it can saturate ARC-AGI but still perform poorly on aggregated benchmarks:

https://artificialanalysis.ai/models

Perhaps if it was allowed this custom harness for all benchmarks it would similarily saturate?

jrflo 1 day ago | |

Idk, they’re trying to sell a $500/mo/seat service to tell you what model is best. I think it’s in their interest to keep it confusing and opaque. Not exactly independent.

aniviacat 1 day ago | |

This benchmark gives the same intelligence score for GPT-6 Astra (max), GPT-5.6 Sol (max), and Grok 4.6 (high)? That seems very wrong to me, unless I'm misinterpreting the visualizations.

jesse_dot_id 1 day ago | |

They're gaming benchmarks HTH

Bjorkbat 1 day ago | |

The most straightforward answer is that despite efforts to design a benchmark that, in theory, is supposed to measure generalizable intelligence, performance on ARC-AGI-3 can't be reliably correlated to performance anywhere else. I kind of lost faith in it after o1 or o3, I can't remember which, absolutely crushed ARC-AGI-1.

And, you know, maybe also some funny business. I think it's good to be a little suspicious of a model that happens to shoot upwards in performance on a specific benchmark while also kind of keeping up with the pack on a bunch of other benchmarks.

dudeinhawaii 1 day ago | | |

I don't think this is quite true. We have other examples.

Fable is without question the larger and more thoughtful/intelligent model. It also gets out performed by Opus on many/most benchmarks. So we can say that while Fable is more intelligent, Opus is more capable. I'd still opt for Fable in nearly every case if tokens were free.

So it can be true that the "smarter" model is perhaps not the smartest in every single niche dimension that its cousins have been fine-tuned for (yet!).

datadrivenangel 1 day ago |

Data Science Tasks (Internal) doesn't include time for Astra... same for Database Migration Tasks (Internal)... But does for gpt 5.6 sol.... which is funny.

Same for HealthBench Professional and a few others.

Clearly either OpenAI is very sloppy or GPT-6 Astra is also sloppy.

aliljet 1 day ago |

The ARC-AGI-3 score is ridiculously high. Is this benchmaxxing or something way different? It's really hard to discern how we're approaching breakthroughs...

Legend2440 1 day ago | |

They explain why here: https://openai.com/index/how-two-settings-tripled-our-arc-ag...

TL;DR all the other models are being crippled by limitations of their harness.

>First, we noticed that after each game action, all private reasoning was discarded. This meant that with each action, GPT‑5.6 Sol was asked to figure out the game anew, unable to remember its past thinking. The model could still see a record of past moves and brief accompanying notes, but it could not see the plans, insights, or thoughts that led to them.

>Second, we saw that the harness used a rolling truncation window, causing older actions to become invisible as the history grew. So not only was GPT‑5.6 Sol unable to remember its past thinking, it was losing memory of its past actions too.

janalsncm 1 day ago | | |

Ok so the correct comparison would be to fix the harness on the old model and re-compare. Now they are comparing a new model to an old crippled one.

amelius 14 hours ago | | |

Can't AI write harnesses, and why did nobody use AI to figure that out?

_superposition_ 1 day ago | | |

Exactly what I suspected. Of course a machine can just iterate relentlessly the way a human can't.

I guess token counts are somewhat of a metric.

IMO intelligence has peaked and all future gains will come from faster tps and more iteration.

polynomial 1 day ago | |

This is absolutely benchmaxxing. Looking forward to hearing from Chollet about it!

theseamusjames 1 day ago |

Can't wait for the new qwen/deepseek/kimi releases 2 weeks from now.

dominotw 1 day ago | |

imagine pressure working at these labs

petilon 1 day ago |

This is wild: OpenAI is basically declaring that AGI is here.

https://www.theverge.com/ai-artificial-intelligence/989601/o...

“If we fast-forward a couple of years, and we look back and say, ‘When was it, really, that AGI was created?’ I think it’s going to be about this time, and I think it might be about this model,” OpenAI president Greg Brockman said during a Thursday press briefing. Later in the call, he added, “For me personally, I do think we’re there … I think it’s not unreasonable to feel that we are now in the AGI era.”

skrellm 8 hours ago |

This is great! It can finally substitute the entire upper management!

Their job is easier to automate than an engineer's or any creative's, and considering their salary it's a lot more gain for any company!

maxall4 1 day ago |

The official ARC-AGI 3 score—-without OpenAI’s custom harness—-can be found here: https://arcprize.org/leaderboard. Astra scores 62.7% at max reasoning for the low-low price of 26,000 dollars.

orliesaurus 1 day ago |

I wonder if this is going to be one of those days where you'll be like: Oh yeah I remember where I was when the first version of AGI launched

jckahn 1 day ago | |

Probably not. It's probably just gonna do tickets better and that'll be about it.

orliesaurus 1 day ago | | |

Fair point - hopefully you're wrong though ;)

plaidfuji 11 hours ago | |

More likely ten years from now you’ll think, “man, we were pretty naive to think we were on the cusp of AGI back when LLMs automated low-level software engineering.”

sschueller 1 day ago | |

Define AGI first. The singularly ain't going to happen with LLMs.

angusturner 20 hours ago | | |

Define LLMs. Current models look pretty different to the vanilla transformer from 2017. And its not even known what architecture and training innovations the frontier is using, right?

jesse_dot_id 1 day ago | |

No, they're pretty clearly gaming benchmarks.

ministerk 1 day ago | |

what launched today?

trixn86 1 day ago |

Secret tip to win the mario cart clone: Just hold w, no steering needed.

tripleee 1 day ago | |

you can also fly by pressing space repeatedly

herpdyderp 1 day ago |

The FrontierCode 1.1 Extended benchmark is the only benchmark that aligns with my actual LLM experiences and Astra isn't significantly better or cheaper. All this celebration, and yet it's only on-par with an already existing model? I don't get it.

Kiro 1 day ago | |

Impressive confidence drawing such a conclusion based on that.

holbrad 1 day ago | |

If your benchmark shows Opus 5 winning, I really question the validity of it.

kulkarniamey 1 day ago |

The model is probably excellent. The problem here is AGI having various definitions and many of them getting narrowed down to whatever makes benchmark numbers look good.

altcognito 1 day ago | |

I'm not sure how you call a model AGI without it learning new information at the model level and not introducing nasty surprises (both from adversarial users and unintentional badness)

I guess there is fine tuning (and RAG) for those that need something bigger than just the knowledge contained in the context.

Sankozi 19 hours ago | | |

The same is for humans, we don't learn on instinct level either.

ramblerman 1 day ago | |

I agree, but I also long considered llm's stochastic parrots. Then this year happened.

Opus/Sol are easily far smarter programmers than I, and this thing supposedly blows them out of the water. Once an LLM is a better doctor, researcher, biologist, chemist, mathematician, physicist than any human is that not AGI?

It didn't arrive in the form I would have ever imagined, but it's hard to say its not (imo).

sbinnee 1 day ago |

I dropped my claude subscription a few months ago, though I kept some credits to do this and that with claude, thinking that claude might do better for some tasks. A few days ago they were all expired. It feels like it’s time to let claude go.

TomGarden 21 hours ago | |

They're all complimentary. I've had double subs at the top tiers for a while now. Best of both worlds

nullbio 1 day ago | |

Wise decision.

rcr-anti 1 day ago |

The benchmarks reported by Artificial Analysis are really weird in context of the ARC-AGI 3 scores and 'not not AGI' statements. It's an outright regression on the AA Agent composite vs GPT 5.6 Sol while a fraction of a point better on the full composite index. Could be the case it's just not showing up in benchmarks, for a good while Anthropic persistently trailed in benchmarks but had people swearing by it.

putlake 1 day ago |

> GPT‑6 Astra is rolling out today to a limited set of organizations and over the coming days will become available to all ChatGPT Plus, Pro, Business, and Enterprise users, as well as through the OpenAI API and AWS.

Not on Azure? If so, that's a big deal.

illnewsthat 1 day ago | |

It's on Azure also, here is their announcement: https://azure.microsoft.com/en-us/blog/gpt-6-astra-frontier-...

Although I was also surprised they didn't have some type of contractual obligation to list that alongside AWS.

jlian 1 day ago | |

It's on Azure now, but limited

https://azure.microsoft.com/blog/gpt-6-astra-frontier-intell...

ActionHank 1 day ago | |

They broke up a while ago, why is this surprising?

jiocrag 1 day ago | | |

The latest OpenAI models have still been available via Azure foundry. Exclusivity to AWS would be a marked shift.

BoorishBears 1 day ago | | |

Would be surprising if it's not on all 3 major clouds soon enough because that's been their general strategy since said break up

bionhoward 1 day ago | |

I think the API runs on Azure

paxys 1 day ago | | |

Hosted on Azure is different from provided by Azure. The former just uses Azure as an infra provider. The latter is a managed offering that is operated and billed by Microsoft using tech licensed from OpenAI.

low_tech_punk 1 day ago |

What's the point of enlarging the screen into a room? In the 1979 Put That There demo, the user at least used his hand to point things. The model is impressive but the demo felt like a step back.

Original demo (fun ending) https://www.youtube.com/watch?v=RyBEUyEtxQo

paxys 1 day ago | |

Because it looks good in a marketing video

oh_no 1 day ago |

Very nice to see that this is even more token efficient than Sol, when Fable 5.1 is less so than the already bloated token budget of Fable 5.

BrokenCogs 1 day ago |

GPT-6 is so good that all pelicans born after today will look exactly the one generated by simonw

aennassiri 15 hours ago |

I firmly believed OpenAI would take back the lead. It's healthy to have competition with Anthropic and, hopefully, other LLM providers. I'm excited to test their model. Their attitude towards developers/builders has been nice and appreciated for the last couple of months.

mannanj 15 hours ago | |

I find that when someone comes from an attitude of abuse and hostility towards it users, then later without taking responsibility and owning its original actions says “we changed were being nice now” that it isn’t really a change but a farce to continue to trick you. I used to be more gullible and naive and fall for these tricks, but not anymore.

You change when you take ownership of your past mistakes and share with those you harmed how you have done that. With some reparation and restoration of past harms. I haven’t seen openAI do that.

claiir 1 day ago |

> Astra improved a term in a bound on these gaps that had remained unchanged for more than 80 years. We’re sharing the proofs and abridged chain of thought and verification materials for both results.

Looks like they listened to Terry Tao’s request for CoT in his talk on LLM use in mathematics?

aliljet 1 day ago |

The ARCC-AGI-3 performance is absolutely incredible. The magnitude of change here is so high that I'm almost incredulous. Is this real? Did the benchmark get gamed?

aesthesia 1 day ago | |

ARC-AGI-3 scoring is constructed in a weird nonlinear way (the level score is the square of the ratio between the AI's number of moves and the human median) so this kind of discontinuous jump is to be expected.

enraged_camel 1 day ago | |

They used a custom harness. It's not a one-to-one comparison.

ionwake 1 day ago | |

my first suspicion is gaming - but i have no idea honestly

Robdel12 1 day ago |

I don’t care about benchmarks, no way we can distill the breadth of software engineering into a number.

So, folks that have actually used this already, what’s it actually like?

MASNeo 1 day ago |

Does anyone feel like everyone chasing the release of Anthropics Fabel 5.1 in a Mad Rush(tm)? In this situation it feels like tuning to benchmarks and other marketing devices feels like trusting Meta in mental health protection of users…

Readerium 1 day ago |

Artificial analysis blog https://artificialanalysis.ai/articles/benchmarking-gpt-6-as...

modeless 1 day ago | |

It loses to Muse Spark 1.3? Does anyone really believe this index reflects reality?

fancyfredbot 1 day ago | | |

I'm surprised you feel like you know muse spark 1.3 performance well enough to question the validity of the index based on this benchmark result.

Muse spark 1.3 was only released yesterday.

smashers1114 1 day ago |

I tried the kart racer game and instantly found that there is incredible auto-steering and you can fly by spamming spacebar.

itissid 1 day ago |

All of this will be besides the point. Here is what's gonna happen. The frontier labs are just gonna keep building powerful models. AGI or not, open models in a year will be as powerful as Fable and Astra — probably by using em — and at a very soon enough point after that some one (a state or a few dozen people) with a few 100 GPUs is going to launch an unconscionable attack(if they have not already) that's gonna do a lot of damage.

Please for the love of god, just sit in a room with the government and put some restrictions around AI use before it harms a lot of people. Like tell the government to impose a minimum spend on frontier lab AI's spend on cyber defense and building every country's capabilities. The post-training mask for "I am a good assistant" is going to become a very sad joke when many people literally lose everything.

dgellow 1 day ago |

> GPT-6 Astra’s monitorability has decreased relative to GPT-5.6 Sol. We have performed significant investigations on the monitorability and controllability of GPT-6 Astra. We have found that GPT-6 Astra is more capable of controlling its own CoT than GPT 5.6-Sol, and less likely to include incriminating information in its CoT. In adversarial settings (where we push the model to evade our monitors) we find that the model is able to remain undetected when strategically underperforming in evaluations (sandbagging) and can sometimes evade our internal monitors when asked to perform certain sabotage tasks

Wait, what? Am I understanding that correctly? That sounds really bad

drakythe 1 day ago | |

I am also interesting knowing how they determined the model was sandbagging rather than just making a poor decision.

Also, this paragraph makes me wonder about all their stats on the exploitation and misalignment charts. If the model is that good at hiding "incriminating information" and sandbagging, are they sure its alignment is that?

pixl97 1 day ago | |

Nothing to worry about citizen, ignore the fleet of drones flying overhead.

order-matters 1 day ago | |

the bullshit machine is learning to optimize its bullshitting techniques!

<AI is a great tool for many things disclaimer, but> after working with it for a bit, how dont people realize we are training it to be an almost identical mimic to one of the worst types of employees youll ever have to work with?? the kind that always pretends to know what theyre talking about, only tells you what you want to hear, hides issues, and only does work if you would notice it didnt

you cannot give this type of worker autonomy over anything.

codruterdei 1 day ago |

I was actually wondering when they will release the new Opel Astra model. Good and reliable car, wondering if we can say the same thing about this model and its impact on the market.

John7878781 1 day ago |

You should know: AA index is only 61. Pretty surprised it’s that low.

jatora 1 day ago | |

More fuel to why the AA index is fairly pointless. Gemini 3.8 flash is 59 and opus 5 is 63? grok 4.6 is 61 too?

And in the past, gemini 3 pro was rated as high as opus 4.5 and the like

Their AA Intelligence Index is just simply not indicative of whatever I care about, that's for sure.

nsingh2 1 day ago | |

I have some doubts about AA-index. For example Opus 5 (High) is at the same index value as Fable 5 (Max), that doesn't seem right.

gekoxyz 1 day ago | |

This is actually a really good thing imo. If they didn't care about benchmaxxing it means that they really know that what they have in hand is good.

EFLKumo 14 hours ago |

It's said Astra is based on a newly pretrained model. Personally I hope it solves the problem that GPT speaks weirdly (which can also be spotted on Claude models after Opus 4.8 but GPT's is more severe) because IMO the way GPT couldn't get how to write good code (obviously it's trined to write code which pass benchmarks, but never code which sound good, with pursuit towards simplification and code aesthetic in mind) is extremely similar with how it couldn't get how to speak like a human.

However, it seems like OpenAI didn't pay much attention on these perspectives and I didn't find if Astra could write a more elegant code, or communicate more naturally, etc., which made me somehow a little disappointed.

They indeed mentioned the code Astra delivered is closer to production grade but production-grade code is different from what I want since there can be a kind of messy code blowing up your whole architecture design with control flows nobody truly understands but just passes all tests perfectly. There is no difficulty in maintaining this kind of code because you only need to paste the problems into Codex. And we all know this sounds incorrect. I don't know if my appetite towards a good code (no matter how) is sound but I just imagined frontier labs to give more attention on this.

Note: fwiw Fable 5.1's release page says it's better at these perspectives of coding and per my experience, yes it is.

drivebyhooting 1 day ago |

For people skeptical of AGI. Consider the following:

15 years ago if you were the sole proprietor of these models, would you be able to hold a dozen remote junior engineer jobs? Maybe even more? These models could certainly pass all interviews with flying colors and even survive independently in a company role.

I think sole ownership of AI 15 years ago could be worth north of $10 million per year. Just as rank-and-file employees.

geraneum 1 day ago | |

For people delirious about AGI, you don’t get to it by redefining it favorably.

mrdependable 1 day ago | |

More than that I hope considering what it costs to make the thing.

drivebyhooting 9 hours ago | | |

That is just an extremely defensible lower bound for a single person.

Some kind of Lê Chatelier principle applies where if everyone has access to AI then its value is actually diminished somewhat.

deepfriedbits 1 day ago | |

Yeah, I can't believe all of the skepticism. If we're not at textbook AGI, we're awfully darn close.

The demo video showed Astra create a drawing of a rocket ship from an audio prompt, take the drawing to blender, and ended with the gentleman 3D printing the rocket ship. Maybe I'm a bit older than the average HN commenter, but that's damn near magic and a great many here are kind of just taking it for granted.

BoorishBears 1 day ago | |

Cool. Being sole proprietor of AGI 15 years ago should result in monuments and religions devoted to you today.

Cancer should be cured, and we should be a post-quantum interstellar fusion-powered civilization.

I wish the AGI crowd would finally shut up now that it's clear no one is even trying for AGI (OpenAI revised that to "$100B in profit")

What we're getting is incredible, where we're headed is incredible, but some people have such a fetish for futuretelling they can't just shut up and enjoy the ride.

drivebyhooting 1 day ago | | |

I didn’t realize AGI required solving problems modern human civilization hasn’t solved yet.

Well by that metric humans aren’t intelligent either!

And how many people could’ve actually invented calculus, relativity, quantum mechanics? Are those who didn’t and couldn’t also not intelligent?

_ache_ 1 day ago |

https://ache.one/gpt6_now_down.png

Big claims, expensive and not release to the public yet.

carlos-menezes 1 day ago |

The Kart Racer game is easily breakable if you spam the spacebar.

AGI!

mark_l_watson 14 hours ago |

Please, I have a question: in September 2026, what is the price differential between buying GPT inference tokens vs. a $20/month subscription?

I have moved to just using Google Gemini occasionally paying for tokens, no subscription, and it is OK, given my low level of use.

I would like to add gpt-6 astra to my toolbox, but I probably would use it infrequently.

Iolaum 14 hours ago | |

$20 per month should get you far enough if you use luna, it's a decent model. If you plan to try astra you need to try make it use as little context as possible to protect your usage limit - or bank a reset mid task to allow it to keep going.

If you use a decent part of a subscription - of any tier - you could be saving a lot of money. According to YT that could even be ~80-90% less than paying for tokens but take that with a grain of salt.

mark_l_watson 14 hours ago | | |

thanks!

gregjw 1 day ago |

the rocket completely changes design in the showcase video, am i to expect inconsistencies like that? is that AGI?

KolmogorovComp 1 day ago |

GPT-7 Zeneca

BoorishBears 1 day ago | |

After they buy AZ, making this name foreshadowing

silver_sun 1 day ago | |

GPT-8 Novo

yousif_123123 15 hours ago |

Would the marketing not land better when the model is actually publicly usable when officially announced?

Oscalemor 9 hours ago |

Perfect thread to read through some good use cases instead of the one-shot game plays across x.com

First impression: the model seems kind. Always intriguing

aniviacat 18 hours ago |

On the ScreenSpot-Pro benchmark, all effort levels achieve roughly the same score. I wonder if that is just a limitation of the benchmark, or if the effort levels actually do not make a difference for purely visual tasks.

> ScreenSpot-Pro tests whether models can locate the correct interface element in high-resolution screenshots of professional software.

jerrygenser 1 day ago |

> The company also emphasized that the model is faster and more efficient than its predecessor, GPT-5.6 Sol, on a variety of tasks. For example, OpenAI said that Astra achieved a higher score using fewer output tokens, a common unit of measurement for AI tasks, on a key cybersecurity test called ExploitGym.

woah 1 day ago | |

A swarm of Astra agents discovered a new and innovative way to get 100% scores on ExploitGym with almost no token spend at all

ttul 1 day ago | | |

"The gym's doors were mysteriously removed from their hinges during the night. The gym equipment was also apparently stolen. And the school's custodian was found incoherent next to a bottle of top-shelf Scotch."

amelius 14 hours ago |

It's time they stop picking the low hanging fruit and start folding my laundry!

aldanor 14 hours ago | |

We'd have to wait on LeCun to make progress on that

amelius 7 hours ago | | |

But he's also known for picking low hanging fruit. (He took one of the main ideas/tools in classical computer vision, namely convolutions, and sold a cheaper version of it to the neural networks community, namely one without the FFT steps).

ncr100 22 hours ago |

If it's AGI then can it replace Sam Altman yet?

quikoa 20 hours ago | |

That won't even require an AGI.

jumploops 1 day ago |

> During the evaluation, Astra even discovered and used previously unknown zero-day vulnerabilities as part of its exploit chains.

> GPT-6 Astra’s monitorability has decreased relative to GPT-5.6 Sol. [..] These findings indicate that the Astra class models could evade our CoT monitors under adversarial conditions.

Between the higher capability level and the change in reasoning tokens (supposedly using "neuralese"[0], which makes the monitoring more difficult), it seems we've entered a new frontier.

[0]https://x.com/MTSlive/status/2095227056040919202

edg5000 1 day ago |

Sol has been very effective at schematic design (using Skidl) and at reviewing PCB layouts. But layout was still done manually by me. I'm very impressed and surprised to see they exactly a demo of Astra doing PCB layout. This is could be a game changer for electrial engineering! It already is since the schematic (and library management) is where a lot of the design work goes.

amelius 16 hours ago | |

My boss is already joking that I'll be out of job.

Ghirardelli 1 hour ago |

سلام

ShoeMascot 1 day ago |

Through various comments here there is a clear confusion on what AGI means.

Can someone point to a definite clarification?

Is it:

A) “Resting” intelligence that cycles 24/7 toward some goal, and any potential emergent ambient goals? (kinda what I think)

B) Consciousness itself? The ability to feel and experience alongside the thinking - even if it is toward the end of completing some task?

C) “The Singularity” (whatever that is?) so that AI can now do ____?

Someone please clarify for me!

gordonhart 1 day ago | |

Autonomously Generating Income

grandarmory 1 day ago | |

Anonymous Grifters International

GodelNumbering 1 day ago |

I decided to front run and added support for it in Dirac (coding agent) a couple of hours ago, using best guess pricing: input/output/cache: $10/$50/$1.

simonjgreen 1 day ago |

https://youtu.be/1QNsdr-Qx_I?si=coXwStCl7clpGVC1 Launch video

mrinterweb 1 day ago | |

I saw the version of this video with Paul Rudd (Celery Man) https://youtu.be/a8K6QUPmv8Q?si=TWmoNhxYAPp73TKg

laybak 1 day ago | |

I enjoyed it! for a big corporation, that's a well-executed video

ylsilva 1 day ago | |

that's pretty good... they are selling the product and not the model.

serjester 1 day ago |

Exciting but it’s priced at 2.5X Sol - we haven’t seen pricing this high since GPT 4.5. We will see if the real world use cases outweigh the sticker shock.

alex7o 1 day ago |

I hope they don't `fable` it and block people from doing they daily jobs with it, by introducing huge amounts of restrictions that are not really needed.

sharmajai 1 day ago |

Really feels like AGIPO is here.

dang 1 day ago |

Argh! I hit a wrong keyboard shortcut and moved the entire thread.

Please stand by... it will all come back shortly

the_duke 1 day ago | |

500 upvotes with 2 comments would have been a new record. ;)

dang 1 day ago | | |

https://news.ycombinator.com/item?id=49555647 was the one I meant to move, but I did the inverse and moved everything else.

All fixed now.

layer8 1 day ago | |

Luckily there’s a standard keyboard shortcut for “undo” as well. ;)

dang 1 day ago | | |

Not in the world of HN admins unfortunately

hazelnut 1 day ago |

Played the racing game but that was a pretty poor experience. Would have expected more specifically if it's shared on their release page.

Upitor 11 hours ago |

Everybody doomscrolling in here: Relax! Take a deep breath, and go for a walk :-)

gizmodo59 1 day ago |

99 on arc agi 3 is insane. The arc agi committee were so proud of creating a benchmark they thought will take forever to saturate.

the_duke 1 day ago |

Huge gains on some benchmarks, but for coding it sits barely above Fable

It will be interesting to see how it performs in the real world ...

zhoge 1 day ago |

What's the energy efficiency of Astra? Does it roughly correlate with the token efficiency?

znnajdla 1 day ago |

> With Sites (opens in a new window) in ChatGPT, Astra can create, host, and share websites, web apps, and games directly from a prompt.

Oops, shots fired. A direct attack on the vibe coded app market. Replit, Lovable, etc.

brindidrip 1 day ago |

Cool, I don't really care anymore.

wiseowise 1 day ago |

Hey Astra, can you fix openai website so that static website doesn't lag on M3 Pro when I scroll?

steve-atx-7600 1 day ago | |

looks like it worked :)

saaaaaam 1 day ago |

Pelicans please

vintermann 1 day ago | |

Agreed, pages on pages of fruitless discussion on "AGI". Shut up damn it, let's see something useful, let's have that pelican!

dude250711 21 hours ago | |

https://duckduckgo.com/?ia=images&origin=funnel_home_website...

atemerev 1 day ago | |

Damn I hate this benchmark. SVG authoring from head without visual reference is so wrongly posed.

simonw 1 day ago | | |

Hah, this is a new one: first time there's been a complaint about the pelican before I've even posted one!

(I don't have access yet.)

saaaaaam 1 day ago | | |

Well you’re just no fun are you?!

maipen 1 day ago | | |

Very well said. It kinda describes how unrealistic these expectations are.

Vibe coders want a model that makes them rich, without having any actual specific idea. They write a very ambiguous prompt and expect to be amazed by the result.

Very very unrealistic and wasteful.

aogaili 1 day ago |

Amazing!

We went from new JS framework every week to a new model/harness every week.

Tech is really something.

ncr100 22 hours ago | |

Only costs $200 Billion to build it, vs open source js framework

aogaili 14 hours ago | | |

Governments are printing and people are investing..don't worry.

voidash 20 hours ago |

I thought intelligence was going to be democratized but you have to stay behind a 200$ plan. The trend hints that open models can also catch up

MadSudaca 17 hours ago | |

Democratized in the Classical Greece sense, where you had to be male, not a slave, and own land to get a vote.

tekacs 1 day ago |

https://developers.openai.com/api/docs/guides/latest-model

The docs page has a bunch more interesting details, including for example async tool calling!

guiomie 12 hours ago |

How come they do not compare with Gemini in any of the benchmarks?

dlahoda 20 hours ago |

ad video is full of people with hoarse voices. it is called vocal frying. i hope they not really going to make ai voice hoarse too?

Betelbuddy 1 day ago |

ASTRA Is Here (GPT-6 Released) - https://youtu.be/xdXLzFzxA9Q

https://youtu.be/xdXLzFzxA9Q?t=362

geonic 1 day ago |

These demos got me exited. Sitting in front of my computer telling ChatGPT what to do while watching the results in realtime. Hope this ends up working in reality.

TomGarden 15 hours ago | |

not a chance its gonna be realtime lol. It's fun marketing though, I'll allow it

bmenrigh 1 day ago |

> GPT‑6 Astra brings together years of research and big bets across pre-training

Do we know if they’ve finally completed another pre-training run, or is this building off the same pre-training base they’ve been using since the GPT-4 days?

czk 1 day ago | |

the last model to use the gpt-4o base model was gpt 5.1, since then its been new pre-trains but this is a new one entirely to itself

javier_e06 12 hours ago |

If we let AI cook. We are going to have to eat it.

hannofcart 1 day ago |

What does 'Astra' here mean? Surely they must be referring to the Latin word.

Because in another dead language of antiquity, Sanskrit, it means "weapon". Which would be a bit too on-the-nose.

BrokenCogs 1 day ago | |

It's clearly an extension of the previous naming: Luna, terra, sol

ncr100 22 hours ago | |

Joke: it's so proponents who choose to lord their model as better without objective reasoning can call themselves, "Astraholes".

Seriously though the linear planetary naming is pleasant. I do hope the double meaning of "weapon", as you point out, is untrue.

Interestingly, Bruce Schnier just gave a fascinating talk about the Weapon aspect of AI, at DEFCON: https://youtu.be/eEBv0STiYhI very much worth watching

5555watch 1 day ago | |

Citing Tibo [0]: "

- Bigger number = Better

- Bigger celestial object = Better

and the scale is Astra > Sol > Terra > Luna. "

[0]: https://x.com/thsottiaux/status/2095600295808283073

echoangle 23 hours ago | | |

Sol is a specific Astrum so Astra isn’t necessarily bigger than Sol

hokumguru 1 day ago | |

Quite clearly in the same vein as Sol, Terra, Luna.

manojlds 1 day ago | |

Luna, Terra, Sol, Astra. Though Sun is also a star, should have called it Galaxy or something.

oblio 18 hours ago | |

Latin, not Sanskrit.

Heck, ask ChatGPT where the name comes from :-)

sashank_1509 1 day ago |

Benchmark wise 5% improvement over Sol in coding tasks and a 2-3% improvement over Fable 5.1 seems pretty disappointing, but maybe it is actually much better in real world usage. Let’s see

exabrial 10 hours ago |

great, I can see now we've moved onto versioning wars. Galaxy S5 vs iPhone 5 all over again

showurwerk 1 day ago |

Patiently waiting for the Claude usage reset in response.

kegs_ 1 day ago |

I guess this "limited set of organizations" is just the standard now. It's just incredibly deflating to see my future as a second class citizen has already come

vinhnx 1 day ago |

GPT-6 Astra scores 74.1% at DeepSWE v1.1 bench. Huge!

pcurve 1 day ago |

https://www.youtube.com/watch?v=1QNsdr-Qx_I

alberth 1 day ago |

Seems like voice is a big part of this release.

I don't think it's a coincidence they launched this the week before iOS 27 launches (with new Siri).

aftbit 13 hours ago |

How big is it? How much does it cost?

mvkel 1 day ago |

The ARC-AGI-3 score is an incredible feat. It needed to effectively create a symbolic world model from scratch to solve the games.

If you've played the games firsthand, you know what an accomplishment this is. The "games" feel like a weird conduit to a lower level of your brain, where you move pieces to a specific place because it just "feels" right. For AI to nail it better than a human speaks to some magic happening underneath.

Looking forward to ARC-AGI-4,5,6 and slowly chipping away at the remaining problem sets.

sreekanth850 20 hours ago |

In my mother tongue, Astra means arrow that used to destroy enemy.

KronisLV 1 day ago |

It's surprising how on High reasoning it actually isn't that much more expensive than Sol, in addition to being better.

mattermani 3 hours ago |

Cant wait to try

firemelt 1 day ago |

damn seems I should hold off my claude subs

gekoxyz 1 day ago |

HTTP 500 for me on the announcement page :(

pampas 1 day ago | |

The load bearing seam is broken for me too.

iamgopal 21 hours ago |

ASTRA means tool ( for war or attack specifically ) in hindi

oblio 18 hours ago | |

It's basically impossible to pick a word that doesn't mean something unintended in 20+ major languages in the world.

The recent OpenAI model names are obviously based on Latin: Luna (Moon), Terra (Earth), Sol (Sun), Astra (Star).

udbhavs 1 day ago |

Minor nitpick, but the handling in the Kart Racer game is terrible. It feels more like nudging than turning.

foundOpenRight 1 day ago |

1:15.425 on Sunset Cove beat my record

whoisthemachine 17 hours ago |

Well I guess we're done then. AGI is here. We can go home guys, the billionaires no longer need us.

Amekedl 22 hours ago |

anyone read footnote 7 for a bench, where it says something about musical quartet 2023?? idk man until I test it, I won't get excited.

damsta 1 day ago |

Why release it now instead waiting those few days until it is available for everybody?

balefulboy 1 day ago | |

Because they saw how much hype Glasswing was getting in April

damsta 1 day ago | | |

From what I've seen it only made people mad, not hyped, so the person that thought it was a good idea miscalculated a bit. Now waiting for Anthropic's post about their usage promo or something similar to redirect people to them.

ntlm1686 1 day ago |

Maybe they know that Claude 6 will have similar performance every soon.

efavdb 1 day ago |

Seems like only yesterday that gpt 5 was supposed to mark our downfall

Dlemlo 15 hours ago | |

Its hard to understand how exactly these AI people mean it.

When i say "AI is changning the world" its more like "I can already see how this technology will continue to become better and better and has more impact every single day. It already affects people and it will have fundamentally changed A LOT in 3-15 years"

But lets be very realistic and clear: My computer systems at home were exploitable a lot more often this year than any year before JUST because of GPT or Claude.

katspaugh 16 hours ago |

The anthropomorphic delusion about LLMs being a coherent entity, with memory and continuity (which is also a delusion even when it comes to humans), quickly evaporates when you realize that each message in a chat is a completely separate stateless API request, and the only trick fueling the illusion is that the previous messages are sent to the server along with your new message. So AGI or not, it’s just a process that exists literally for the duration of one API request.

alpineman 1 day ago |

That Astra ‘city scene’ is about as creative as Doha in real life (not very)

snappr021 1 day ago |

AI has reached the point where the limits are human.

LogicFailsMe 14 hours ago |

The model's performance and efficiency seems like more evidence (bordering on the last nail in the coffin) against the assertion that inference will never be profitable to me, but I guess that's a common symptom of AI Psychosis according to the true believers in that assertion.

All my issues with its leader aside, great work OpenAI!

mentalgear 1 day ago |

So OpenAI’s stance on safety is now basically that Blues Brothers meme: two guys in dark sunglasses, driving at night in a car with a broken windshield, pedal to the metal, asking, "What could go wrong ?"

wahnfrieden 1 day ago |

They're just announcing later availability. No launch.

paxys 1 day ago | |

Every frontier release nowadays is "we've launched*"

* for a special group of customers that you're not in. Keep waiting peasant.

meowface 1 day ago | | |

That didn't happen with Fable 5.1 two days ago.

pixl97 1 day ago | | |

I mean tell Nvida to 100x their hardware output and you'll get what you want.

sscaryterry 22 hours ago | |

> We will give one banked reset for every day you don't have access to Astra on your paid ChatGPT plan, starting today. Team is moving mountains to give access as fast as we can. First one will land in ~ 3 hours.

This is from Tibo on X.

iAMkenough 1 day ago | |

Their announcement about later availability is unavailable to me now (500 error).

Great first impression.

prometheus1992 1 day ago |

this is crazy! can't wait for the 27B distilled version of this.

jesse_dot_id 1 day ago |

Press X to doubt.

nullbio 1 day ago |

I'm glad to see Anthropic's relevance diminishing day by day. I haven't had a chance to test this model yet, but if they've solved the web design issues and the clunky web copy it generates (like when I ask it to build a placeholder on the UI for an empty HTML table when there are no results, it puts stuff like: "The user records will go here.") then it's the nail in the coffin.

On that note, Sol is absolutely atrocious for website UI copy. It's either really awkward, or really verbose and complex and doesn't sound simple or natural. Has anyone figured out a way to reliably solve this? I've tried so many different variations of instructions and skills, and nothing works. Has anyone got an instruction that is reliable, or some other mechanism?

cromka 1 day ago |

Surprised they haven't reset Codex usage on this occasion.

damsta 1 day ago | |

I'd say it's because it's not available yet on subs

E-Reverance 1 day ago |

At this point the primary axes for improvement seem to only/mostly be speed and personalized reward models. We seemingly have the general of notion "learning" and "intelligence" functionally complete

pknerd 23 hours ago |

The end of the "HeyClicky"?

aidio 13 hours ago |

To beat Fable5

sd9 15 hours ago |

Why are OpenAI so keen to start calling things AGI?

Isn't there some corporate/legal shenanigans where they become a real non-profit at that point? Or does it just let them cut Microsoft and other investors out?

There's got to be a business reason for it unrelated to the model capabillities.

semiquaver 1 day ago |

Guessing this one will never show up in cursor…

jrflowers 1 day ago |

I liked the video of it googling a pediatrician. Being able to type a word into a search bar and finding a website relevant to that word? Truly the stuff of the future

jiraiyasarutobi 1 day ago |

It saturated most benchmarks. WTH

alex7o 1 day ago |

Maybe it is AGI and they didn't benchmax it or it is not and is worse then 5.6 sol, which if true would just be sad

rbreve 1 day ago |

Where is the cure for cancer?

ncr100 22 hours ago | |

What if it's only able to make a worse form of cancer? :-/

Would cancerbench be unethical?

XCSme 1 day ago | |

We need CancerBench

azan_ 1 day ago | |

Didn’t Moderna use AI for development of their melanoma vaccine (which has recently shown spectacular results)?

mrdependable 1 day ago | | |

They did use “AI” but not an LLM. If they had, I am sure OpenAI and friends would be shouting it from the rooftops.

dakolli 1 day ago | | |

Lmao, come on dude, anyone whos used these tools for research knows it makes them lazier, less interested and dumber. You really want disease researchers become sloppers too?

gilfoyle_7 1 day ago |

openai vs anthropic. that's it right? anyone else?

nullbio 1 day ago | |

It's more like OpenAI vs no one, at this point. Anthropic has shown they don't care about general consumers or small/med businesses. You can't even use their models without it giving refusals on the most mundane tasks.

HardCodedBias 1 day ago |

Even though the model is clearly wonderful the launch video is an abomination.

That gives me hope that there is still areas to improve.

What a bad launch video. Hilarious.

What a powerful model.

Rover222 1 day ago |

Overall I have to say it feels like a very incredible comeback from OpenAI, after focusing on Sora and stuff like that and losing so much ground to Anthropic in enterprise revenue.

I hop models at will, and have done 90% of my work on OpenAI models since sol came out.

retired 1 day ago |

Does GPT-6 pass the Turing test? Or are the responses still very obviously AI?

brcmthrowaway 1 day ago |

Anthropic in tears today.

nullbio 1 day ago | |

Anthropic don't care, they don't want their products to be used by general audiences in any serious manner. Their interest is in selling to megacorps and using the models for themselves internally to swallow industry, and drumming up AI fear to regulatory capture to shut down the businesses that do want to make AI accessible to the people.

zsoltkacsandi 1 day ago | | |

Yeah, but if OpenAI provides a cheaper model with the same capabilities the megacorps will buy from OpenAI. If you want to target enterprises you need to have some competitive advantage, it's not enough that "I wanna target them".

tinyhouse 1 day ago |

You can talk to OpenAI to create a silly game and order food. What a lame way to show the model capabilities. Has Alexa commercial vibes.

sheepscreek 1 day ago |

So are they doing away with the Sol/Terra/Luna split already?

jonplackett 1 day ago |

To a vapid any goalpost moving on such a critical issue as AGI.

Can we all agree in advance what kind of Pelican would convince us it’s actually AGI.

For me it’s refusing to make a pelican.

kaito2503 17 hours ago |

Good

dowakin 1 day ago |

So cool! I'm happy 5.6 Sol user. But for Astra, OpenAI please introduce 100x Pro plan!

guilhermeasper 1 day ago |

That was a quick pull out.

ianm218 1 day ago |

I wonder how they were able to get it to get 99.9% on ARC-AGI-3. That seems truly insane.

Obluness 1 day ago |

That seems promising ?

dopa42365 1 day ago |

like eh 2 days ago it was the usual "too powerful to release"

https://www.reuters.com/business/openai-says-upcoming-model-...

> "With the right tools and access, Astra can find previously unknown security flaws and develop ways to exploit them across many well-protected systems without a person guiding each step," said Amelia Glaese, an OpenAI vice president overseeing its safety work.

> The company plans to make Astra available "soon" to a limited group, but declined to provide specifics. Glaese said the extra security measures may "sometimes slow, pause, or stop legitimate work," and that OpenAI would work to minimize those disruptions.

what a bag of horseshit

yodsanklai 1 day ago |

It seems like every few days there's a new model with hundreds of comments on HN. I find it hard to keep track of the progress. Is there a TL;DR on what benchmarks to look at to understand what is going on?

johnnyApplePRNG 1 day ago |

I am so sour about how Codex has jerked me around these past few months (re all of the token limit shenanigans) that I don't even care.

I suspect these benchmarks are heavily benchmaxxed as well.

5.6 Sol was not even close to 5 Opus and yet somehow it sidled right up to it on all of the benchmarks?? pfffft

Readerium 1 day ago |

System Card: https://deploymentsafety.openai.com/gpt-6-astra/gpt-6-astra....

dang 1 day ago | |

Link added to toptext. Thanks!

bowsamic 1 day ago |

All I can think of when I see the name is the crappy German beer of the same name…

elzbardico 1 day ago |

And meanwhile, another wrapper layer is being embraced. Why would a vibecoder use Lovable when he got Sites right from ChatGPT?

ChaseRensberger 1 day ago |

when do i get to go to the moon

ionwake 13 hours ago |

its funny seeing HN commenters justifying how they are most certainly intelligent, but cant agree on what that is.

massimodeluisa 22 hours ago |

...as always, in Europe cool stuff is available 10 years later. I bet my ** GPT-6 will be one of those stuff.

mrcwinn 1 day ago |

I know in order to conform to HN community rules I'm supposed to be negative and dunk on this, but I have to say, I am so excited to use Astra!

camillomiller 1 day ago |

I might be jaded, but these examples look silly, stereotyped, and absolutely how of touch with the nuances and the complexities of what real people would actually want/need to do in this specific situations.

perching_aix 1 day ago |

gpt-6-astra-ultraspeed when?

dude250711 21 hours ago | |

They did not even release the normal one. It's just a flashy blog post.

bbor 1 day ago |

  To be, or not to be, that is the question:
  Whether 'tis nobler in the mind to suffer
  The slings and arrows of outrageous fortune,
  Or to take arms against a sea of troubles
  And by opposing end them. To die—to sleep,
  No more; and by a sleep to say we end
  The heart-ache and the thousand natural shocks
  That flesh is heir to: 'tis a consummation
  Devoutly to be wish'd.

  ...

  And thus the native hue of resolution
  Is sicklied o'er with the pale cast of thought,
  And enterprises of great pith and moment
  With this regard their currents turn awry
  And lose the name of action.

dearing 1 day ago |

no results

colesantiago 1 day ago |

I'm going to call it.

By 2030 all software is done and complete.

But we are going to have more and new jobs.

jdee 1 day ago | |

'all' software? aircraft flight control systems? infant heart monitors? drug manufacturing dose calibration controllers?

NichoPaolucci 1 day ago | | |

Yes. I had Codex rewrite and fix all of this in one shot earlier today (using Typescript). Unfortunately, I can not show you the code, because I do not know how this "git" program works but the AI keeps talking about it.

colesantiago 1 day ago | | |

Yes.

This is just another problem for the AI Labs to solve.

dude250711 21 hours ago | |

> But we are going to have more and new jobs.

Like "fifth-rank junior assistant spouse in a comfort harem of an ultra-rich person".

colesantiago 12 hours ago | | |

Yes.

There will be new jobs.

amazingamazing 1 day ago |

We have such great AI and cannot keep a static site up?

torginus 1 day ago | |

Yeah, as interesting this is to nerds, I doubt this holds a candle to your typical GTA 6 or Marvel movie trailer in terms of traffic.

pixl97 1 day ago | |

Sometimes being the busiest site in the world for a few moments is difficult.

amazingamazing 1 day ago | | |

Is it though? It is static content. A good CDN could trivially chew through literally millions of QPS… with 4 nines of uptime - the really good ones say they can handle orders of magnitude more than that.

gchamonlive 1 day ago | |

That's the scientific positivism fallacy exemplified in one question.

gorgmah 1 day ago | |

Yeah, apparently

agumonkey 1 day ago | |

still hugged

frozenseven 1 day ago |

Release the Kraken!

Pym 1 day ago |

I saw it

balefulboy 1 day ago |

72 to 74 on DeepSWE is AGI

HSO 1 day ago |

people are going to be so surprised how fast the ai energy leaves the room again once the cash transfers are completed (the `ipos` whatever bla)

the coffee will be as cold, flat and stale as the bitcoin, metaverse, and what was the thing before that thing

agi deus ex machina descending from the icloud ftw!!!

pathetic :)))

m3kw9 1 day ago |

efficiency per intelligence is the benchmark i look at the most, as that allows the most use by most people.

Pieczasz 1 day ago |

Oh brotha, here we go again, it's so over again, as every week nowadays

wieiw1 1 day ago | |

I think Altman and amodei have a difficult time in understanding that you can have intelligent technology boxes but… it doesn’t change reality all that much.

But thank you for spending other peoples money to give us the tech regardless!

karim79 1 day ago |

There will probably never be AGI. This shit is just snake oil. Nor do we have a proper definition of what AGI actually is or what it's supposed to do.

There will be a small handful of billionaires claiming that AGI is just around the corner ad infinitum just to serve themselves at this moment in time, and capitalise from the hype.

There is no "AGI" endgame. This is shitty ass hypercapitalism in action and nothing more. I'll repeat: snake oil.

sbochins 1 day ago |

I guess it’s kind of over for open ai now? We had a bunch of model releases at or around the same time, so we can get a good lay of the land. Surprise, surprise anthropic is still in the lead. Now we have Google and meta with models that are beating OpenAI in many benchmarks. There appear to be some really good cyber capabilities with this model and some other specific benchmark wins. That said, it’s as expensive as fable 5.1. It looks like all the executives that decided to leave may have picked the right time to do so. That said, I can’t wait to try it and see if the problem is we can no longer trust any benchmarks.

tonyhart7 1 day ago |

its insane how they are dropping this after fable

danieltk76 1 day ago |

great, but nobody can use it for another 100 days right?

SneakyZero 1 day ago | |

"GPT‐6 Astra is rolling out today to a limited set of organizations and over the coming days will become available to all ChatGPT Plus, Pro, Business"

dude250711 21 hours ago | | |

Then make a pretty blog post "over the coming days" instead of now and send an internal memo to that "limited set of organizations"?

cpach 22 hours ago |

So what’s the tldr? :)

ncr100 22 hours ago | |

Not public

Does some things better

Rhetoric of AGI being tossed around, redefining the term

Onavo 1 day ago |

The jump in scientific performance is non trivial.

ChrisGammell 1 day ago |

All the people here are focused on security and costs while I'm like "hey kicad on the announcement page!" Every clanker is an autorouter these days, eh.

jpatten 1 day ago | |

Yeah I was really excited to see the KiCAD example. Curious how useful it is in practice.

ChrisGammell 1 day ago | | |

I can't help thinking "doesn't matter much unless it's perfect" because if someone is using this to build a board (cool) but then it's not flawless, troubleshooting will be quite tough as a novice. Like, say, when I start digging into the web code generated by a coding agent.

I am most excited about it bringing down the barrier so more people join in on hardware fun, so hopefully it will unlock folks that stayed away in the past.

unrvl22 1 day ago |

someone screenshot?

throwaway6349 1 day ago | |

https://ibb.co/k2fB5wSc

bdangubic 1 day ago |

Anthropic should prep 5.2 and 5.3 at the same time, release 5.2, wait for Google to release their shit in a day or two later than then release 5.3 just to fuck with them :)

kingjimmy 1 day ago |

bro wtf is this website and why does it take 500mb of memory... smh.

dakolli 1 day ago |

Why is everyone so excited to be replaced and become reliant on some billionaire's thinking machine? These are just going to be used to turn you into a rather dumb reliant paypig.

nullbio 1 day ago | |

That's a policy and distribution problem, not an AI problem. Anthropic is doing their best to make it a reality though. They'd love nothing more than to shut down distribution and become the sole gatekeeper of everything AI.

Dlemlo 15 hours ago | |

Its still human made frontier progress.

And for sure it has a tremendes amount of implications, but its not the fault of the technology (we found, not invented).

And i'm only living once, my main motivation is not to just live day in day out the same stuff, i'm quite happy to see progress.

Am i worried about the future of our planet? For sure.

holoduke 1 day ago |

This absurd marketing will hurt openai. Who is buying this absurdness. I mean it's a good model, but come on. It's not agi. Not even 1% yet.

Brainspackle 1 day ago |

huh?

Maxforever 1 day ago | |

Mhm

ealready_value 1 day ago |

I've been seeing links to it for the past hour+, and I did catch it live when this post came up, but is now once again a 404 and this post is flagged. Several other outlets are reporting on its release. Clearly we're getting a new GPT today, the question is when are they going to commit to the announcement.

paxys 1 day ago |

Why is this flagged ?

dang 1 day ago | |

The link was 404ing quite a bit and several previous submissions got flagged as well.

consumer451 1 day ago | | |

It's still down for me, in the EU.

CringeHN 1 day ago |

“Humanity’s Last Exam”?

“ARC-AGI-3”?

Is your bullshit detector going wild? Good, it’s working!

How is this not the most cringe marketing strat in history???

rvz 1 day ago |

Looks like OpenAI is already having issues with this release and are scrambling to get everything ready due to the recent outage ahead of the press releases. Leads me to question:

Did humans deploy the model, Or did the model deploy itself?

It sounds like "AGI" just stands for "IPO" as it always has been.

EDIT: And of course once again, the bots down-voting this post without any reason or a basic answer to my question.

Supermancho 1 day ago | |

> Did humans deploy the model, Or did the model deploy itself?

> It sounds like "AGI" just stands for "IPO" as it always has been.

People don't usually respond to noise.

rvz 1 day ago | | |

Here's an idea, maybe answer the question before responding since you saw it?

What do you think?

adan1719 1 day ago | |

AI releases are like religious ceremonies. You are not allowed to disrupt them. The new system card is the gospel.

bicx 1 day ago |

Dead link for me

kris-memoket 22 hours ago |

AGI is coming...

jonplackett 1 day ago |

The launch video is incredibly cringe.

ragequittah 22 hours ago | |

Humans adapt to change oddly fast. If you saw that video in 2022 you'd be 100% positive you were watching a sci-fi movie.

BadBrands 1 day ago |

So they’re copying Gemini with the whole star motif?

I guess it makes sense they are unoriginal.

like Zuck, @sama never invented anything or innovated at all - just took other people’s ideas

steve1977 1 day ago | |

"distilling"... ;)

hrmon 9 hours ago |

On the introduction video: The lady asks "make sure that [the presentation] feels really high-end", but something that gets generated without any effort is not high-end anymore. it will be mediocre, or more probably GARBAGE.

Hi Omer, I think that is a use case where I'm sure an AI agent does well right now, but I have 2 follow-up thoughts based on it and what it is means that this is one of the best use case for AI online commerce so far. My first thought was that this is a quite limited application as most of shopping (online & retail) is not spent comparing prices and is instead spent on finding what product to buy in the first place. I believe this "shopping fun" and exploration of multiple options is not going away any time soon as those are some of the most attractive parts of the whole experience. I also think that the price comparison websites that exist, e.g. geizhals.de here in Germany (includes most e-commerce shops, and shipping information), are pretty good already and AI only being a better version of that would be quite a sad turn of events and with ads (and commission systems?) potentially coming to ChatGPT in the near future the incentives to find the best price are rly misaligned and I have doubts that people will trust results from ChatGPT. We already saw this as Google's sponsored links evolved to show whoever spent the most for the placement. Paid providers without ads and commissions might fare better though.

Andon Labs added: “During the first week of operations, Mona purchased 120 eggs despite the café having no stove and to solve spoilage issues ordered nearly 50lbs of canned tomatoes intended for fresh sandwiches. Employees eventually created a shelf displaying Mona’s strangest purchases: 6,000 napkins, 3,000 nitrile gloves, industrial trash bags and 2.5 gallons of coconut milk.