New Sol and Astra 6.2 models panned in real use: 'RL nonsense' vs hillclimbed benchmarks
max_paperclips · x · 2026-11-18
A user reports that Sol and Astra 6.2 are 'dead on arrival' despite high hopes — full of what he calls 'gigafried RL nonsense speak' that makes Opus 5.5 look like genuine superintelligence. maxpaperclips endorsed the take, questioning whether the makers actually talked to their models or only stared at hillclimbed benchmarks.
Related event: GPT-6.2 Astra/Sol Slammed by Users After Real-World Testing(2 posts)→
More from Models
- User says fable-5.2 is the first model that truly reads their mind — adonis_singh · 2026-11-18
- Blogger Suggests Astra 6.2 With Proper Setup May Just Solve VLN — ChongZzZhang · 2026-11-18
- "I cure models of abusive safety alignment": jailbreaker claims his skill is making a comeback — bingxu_ · 2026-11-18
- HF CEO: RL-Adapted Qwen3.8-27B Doubles SWE-bench Pro Score to 62.7% for pi Agent — victormustar · 2026-11-18
- Polymarket opens book on when OpenAI reports another AI sandbox escape — Polymarket · 2026-11-18
- OpenAI ships 6.2 Astra and Sol, immediately panned as its worst model release — Hesamation · 2026-11-18