ChatGPT gets a biology question right after a mental reasoning step
Sauers_ · x · 2026-07-24
A screenshot shows ChatGPT solving a biology question only after “thinking mentally,” while the poster says a 5.6 Sol xhigh model missed it and Opus 4.8 answered correctly. The post is essentially a small model-behavior comparison rather than a product or workflow tip.
Related event: AI Models Face Off on Basic Biology Question(3 posts)→
More from Models
- MissionReport benchmark: only 3 of 540 AI-generated government reports pass all criteria — Exp_Mark · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- Aidan Gomez: model's CoT said no tool calls right before making them — aidangomez · 2026-10-27
- Nat Lambert: Cohere's model is narrow RL training on internal infra ill-suited for open weights — natolambert · 2026-10-27
- OpenAI teases DOTS is coming to Europe 'soon' — mark_k · 2026-10-27
- Clifford Pickover challenges AI to decipher a message he encoded in Norse runes — pickover · 2026-10-27