Are Local LLMs Good Enough for Vibe Coding? Gemma4-26B-A4B vs. Qwen3.5-35B-A3B | Dark Hacker News