Third-party 10-dim eval shows DeepSeek V4.1 big reliability gains over V4-Pro-0813
teortaxesTex · x · 2026-09-13
Blogger @servasyyai ran DeepSeek V4.1 Flash through a custom ten-dimension eval, comparing it with V4-Pro-0813. V4.1 shows markedly improved reliability, posting a modestly higher second-round pass rate, while V4-Pro collapsed due to scoring zero on coding. DeepSeek had claimed at launch that V4.1 Flash beats Claude Fable 5.1 and trails only GPT 6.0 Astra; this third-party run offers an independent check on those claims.
More from Models
- Leak: Google Reportedly Has Another 'Hella Chonky' Model After Carbon — iruletheworldmo · 2026-10-30
- Users report across-the-board GPT quality drop: weaker writing and broken instruction following — Kyrannio · 2026-10-30
- OpenAI retiring waves of legacy models: GPT-5 family dies Dec 11, GPT-4.1 by Jan 2027 — ThePeterMick · 2026-10-30
- DeepMind CEO Demis Hassabis tells Musk: testing with trusted partners, coming soon — demishassabis · 2026-10-30
- Adobe stands firm on not training on customer data, using licensed stock instead — omooretweets · 2026-10-30
- Early testers vouch for Mistral's new model: fast, cheap and solid on real tasks — _AndrewZhao · 2026-10-30