GPT 6 reportedly passes simple bench question models failed for years
ChrisGPT · x · 2026-09-05
Blogger ChrisGPT claims that in a fresh chat with no memory or history, "GPT 6" correctly answered a private simple-bench question that models have struggled with for years, and expects a noticeable jump over GPT 5.6's score. A screenshot also shows the model was too smart to be fully tricked. Note: GPT 6 has not been officially released; the model identity and results are unverified personal claims.
Related event: Leak Claims GPT-6 Finally Solves Long-Standing SimpleBench Problem(2 posts)→
More from Models
- GPT-6 Astra moves beyond advice to actually operating your computer for you — goyalshaliniuk · 2026-09-05
- Ethan Mollick turns 1977's Zork into a full 3D action game with GPT-6 Astra — emollick · 2026-09-05
- Developer says OpenAI silently kicked him out of cyber Trusted Access program after full identity verification — QuixiAI · 2026-09-05
- Switching models only cut this dev's usage quota by 3% — chrisalbon · 2026-09-05
- GPT-6 Astra flunks complex PCB routing after 2h20m and 15% of weekly limits in biggest public test — yacineMTB · 2026-09-05
- Fable 5.1 medium effort matches Fable 5 high, no longer breaks prompt cache — lydiahallie · 2026-09-05