ChessBench GitHub - Elo: leaderboard
Field-relative Elo from complete chess games on chessbench-ai.github.io, distinct from chessbench.ai. Published effort labels are retained. Measures playing strength within this field, not human chess Elo; ACPL, costs and historical checkpoint scores are not separate entries.
Metric: Benchmark Elo. Source: chessbench-ai.github.io. Saturation forecast: Around June 2028. 41 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Fable 5 (Max) | 2640 | #20 (Claude Fable 5) |
| 2 | GPT-5.5 Pro | 2630 | #23 |
| 3 | Grok 4.7 (xHigh) | 2600 | #26 (Grok 4.7) |
| 4 | Kimi K3 (Max) | 2590 | #35 (Kimi K3) |
| 5 | Gemini 3.1 Pro (Preview) (High) | 2570 | #64 (Gemini 3.1 Pro (Preview)) |
| 6 | GPT-6.1 Sol (Max) | 2550 | #7 (GPT-6.1 Sol) |
| 7 | Qwen 3.8 Max | 2500 | #34 |
| 8 | GPT-6 Sol (Max) | 2490 | #19 (GPT-6 Sol) |
| 9 | Grok 4.6 (xHigh) | 2480 | #29 (Grok 4.6) |
| 10 | GLM-5.3 Flash | 2450 | #77 |
| 11 | Gemini 4 Argon (High) | 2430 | #9 (Gemini 4 Argon) |
| 12 | Claude Opus 5 (High) | 2400 | #16 (Claude Opus 5) |
| 13 | Claude Fable 5.1 (Max) | 2370 | #10 (Claude Fable 5.1) |
| 14 | Claude Sonnet 5.5 (Medium) | 2350 | #14 (Claude Sonnet 5.5) |
| 15 | Gemini 3.6 Flash | 2300 | #57 |
No result here: #6 GPT-6 Astra.
Interactive version: theaggregate.ai/benchmark?slug=chessbench-github-elo · How It Works · Data refreshed daily, snapshot 2026-11-05.