Community evals put GPT-6 Astra (max) at 45, ahead of Fable 5.1 and Opus 5.5
PawelHuryn · x · 2026-09-23
Blogger PawelHuryn is running repeated-sample evals across frontier models:
- GPT-6 Astra (max): 45 (n=3), currently on top
- Fable 5.1 (max): 43 (n=1)
- GPT-5.6 (max): 42.5 (n=2)
- Opus 5.5 (max): 41.7 (n=3)
- Opus 5.5 shows a steep effort-level gap: xhigh 36, high 31.7, medium 30.3
Notably, GPT-6 Sol at n=1 scored only 32 on max — below the median for Muse Spark 1.3 (max) — and more runs are planned to confirm. Next up: Opus 5.5 (low) and more GPT-6 Sol effort levels. Model names are unverified; treat figures as preliminary.
Related event: Bug Hunt Bench: GPT-6 Astra Tops 105-Bug Fixing Test, MiMo Stuns on Value(18 posts)→
More from Models
- YC-backed Monte launches post-training platform to build proprietary models from your agents' real work — ycombinator · 2026-11-12
- Leak Claims OpenAI Will Launch Entire GPT-6.2 Family Tomorrow to Counter Opus 5.5 and Fable 5.2 — heypearlai · 2026-11-12
- ARC-AGI-3 leaderboard: models score 90%+ but at $10K+ per task — haider1 · 2026-11-12
- LMArena Launches Agent Arena for Head-to-Head Agent Evaluation — arena · 2026-11-12
- LLMs lack hyperlocal municipal data, explains disputed Texas power outage debate — csuwildcat · 2026-11-12
- Claude API's per-message effort lets you switch reasoning levels without breaking prompt cache — lyraxana · 2026-11-12