Hands-on: GPT-6 Sol xhigh feels worse than 5.6 Sol Max at coding
power97992 · reddit · 2026-09-23
A user reports that GPT-6 Sol at xhigh effort feels worse than 5.6 Sol Max, suspecting 6 Sol loses on certain coding tasks even at max effort; DeepSWE evaluations agree. Cheaper, but apparently a regression.
More from Models
- Xiaomi launches open-source MiMo-V2.6 Pro & Flash, tops AI index at 46 among open models — burny_tech · 2026-09-23
- OfirPress: benchmark gap comes from a 166-task subset and pass-rate vs full-completion metrics — OfirPress · 2026-09-23
- Cline claims MiMo-v2.6-Pro is now the top open-weights model on AI index — burny_tech · 2026-09-23
- 'Sorry about that model': Opus 5.5 joke thread hails a huge jump over Opus 5 — repligate · 2026-09-23
- Claude Opus 5.5 wows users with creative output: 'draw everything you want' — repligate · 2026-09-23
- Ofir Press explains why third-party agent benchmark scores look higher: subset + different metric — OfirPress · 2026-09-23