AI Continuous Learning Test: Opus Gains Elo Over 200 Chess Games While GPT Regresses
Peter Gostev ran a 200-game chess continuous-learning benchmark where models play Stockfish using only notes, no weight changes or engine help. Claude Opus improved from 1500 to 1760 Elo, while GPT-6 Astra regressed from 1810 to 1400.
2026-10-04 ~ 2026-10-05 · 3 related posts
- Continuous learning benchmark has models learn chess over 200 games — Elo barely improves — imjustnewatai · 2026-10-04
- Opus climbs from ~1500 to ~1750 Elo learning chess while Astra regresses — inductionheads · 2026-10-04
- Claude Opus gains 250 Elo playing 200 chess games, while GPT-6 Astra drops 410 — 新智元 · 2026-10-05