GPT-6.2 Astra regresses heavily on SimpleBench, ranking 14th-15th
PowerSausage · reddit · 2026-11-19
- A Reddit post shows GPT-6.2 Astra (77.3%) and Astra Pro (79.2%) regressing sharply on SimpleBench, landing 14th and 15th — below the human baseline of 83.7% and far behind GPT-6 Astra Pro (86.5%).
- Claude Fable 5.2 tops the leaderboard at 94.9%, followed by Gemini 4 Carbon (92.6%) and Gemini 4 Argon (91.7%); Claude Opus 5.5 sits fourth at 88.4%.
Related event: GPT-6.2 Astra regresses sharply on SimpleBench(2 posts)→
More from Models
- Bug Hunt Bench: GPT-6.2 Astra max tops at 62.7, beats Fable 5.2 at half the cost — PawelHuryn · 2026-11-19
- Dev observes Google's model launch pace has dramatically sped up since Argon — cgarciae88 · 2026-11-19
- Amazon's in-house web search agent takes #1 on Search Arena with score 1266 — arena · 2026-11-19
- UI screenshot appears to show unreleased 'GPT-6.2 Sol' with reasoning-effort slider — paw_lean · 2026-11-19
- Kaggle lays out roadmap to fix AI benchmarks: relevance, saturation, verification, infrastructure — MeganRisdal · 2026-11-19
- Mistral Large 4 coding test: 2 hours of agentic work burned just €0.82 of quota — qtnx_ · 2026-11-19