Test: Coding Model Quality Dropped After Restoration
wightmanr · x · 2026-07-05
The author shares a comparison of a coding model: before it was restricted, it completed a nontrivial feature in half a day with comprehensive coverage, and self-check and Codex review found no obvious issues. But after restoration, it returned to the usual back-and-forth tuning state, with omissions similar to Opus 4.7/4.8 or Codex (GPT-5.5 xhigh), no longer a significant leap.
More from Models
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11
- GPT-5.6 writes well but is instantly forgettable, user complains — BasedRaddka · 2026-09-11
- Opus Refuses Protein Research Codebase Over 'Safety' Concerns, Dev Considers Rolling His Own — josephdviviano · 2026-09-11
- User Hails Unconfirmed 'DeepSeek 4.1 Flash' as an Inflection Point in LLMs — himanshustwts · 2026-09-11
- Terminal Bench v4: GLM-5.3 Leads at 41.9%, Kimi-K3 Underwhelms at 12.6% — Ok_Warning2146 · 2026-09-11
- GPT-6 Astra beats Factorio with enemies in 44 in-game hours at ~$4,500 API cost — liminal_bardo · 2026-09-11