AI Chess Leaderboard (Continuation): leaderboard
LLMs continue chess games from mid-game positions. Ranked by Elo rating derived from head-to-head matches across 211 evaluated models. Tests tactical calculation and endgame play from non-standard starting states.
Metric: Elo. Source: dubesor.de. Status: years away from saturation. 265 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 1810 |
| 2 | GPT-4.5 (Preview) | 1800 |
| 3 | Gemini 3.8 Flash | 1772 |
| 4 | Gemini 3.7 Flash | 1736 |
| 5 | Gemini 3 Flash (Preview) | 1720 |
| 6 | Gemini 3.1 Pro (Preview) | 1700 |
| 7 | Gemini 3.6 Flash | 1695 |
| 8 | Gemini 3.5 Flash | 1685 |
| 9 | GPT-4o ChatGPT | 1584 |
| 10 | Gemini 3.5 Flash Lite | 1522 |
| 11 | DeepSeek V4 Pro (0813) | 1502 |
| 12 | GPT-4o | 1493 |
| 13 | GPT-5 | 1471 |
| 14 | GPT-5 Chat | 1465 |
| 15 | GPT-5.5 | 1462 |
Interactive version: theaggregate.ai/benchmark?slug=ai-chess-leaderboard-continuation · How It Works · Data refreshed daily, snapshot 2026-09-05.