Mistral Large, Claude 3.5, GPT-4o, Gemini 1.5 Stack Up in Addition Accuracy | Dark Hacker News