ChessBench Elo
Games · 2026-07-28
ChessBench (chessbench.ai) has language models play full chess games against reference opponents and scores the moves; it is unrelated to the chess-bench.com puzzle benchmark carried as `chessbench`. Monte Carlo Elo ratings anchored to fixed reference players spanning the strength range. An unbounded comparison-derived scale, so it carries no declared bounds and does not enter the score-level capability battery.
Top models (higher is better)
| Model | Score |
|---|---|
| GPT-6 Astra | 1769 |
| GPT-6.1 Sol | 1557 |
| Gemini 3.8 Flash | 1527 |
| Gemini 4 Argon | 1485 |
| GPT-5.5 | 1343 |
| Claude Opus 5.5 | 1342 |
| GPT-6 Sol | 1300 |
| Gemini 3.7 Flash | 1273 |
| GPT-5.6 Sol | 1213 |
| Muse Spark 1.3 | 1206 |
| GPT-5.1 | 1203 |
| Gemini 3.6 Flash | 1189 |
| Muse Spark 1.1 | 1153 |
| Gemini 3.1 Pro Preview | 1147 |
| Muse Spark 1.2 | 1126 |
| GPT-5.6 Luna | 1101 |
| Gemini 3.5 Flash | 1094 |
| Claude Fable 5.1 | 1092 |
| GPT-5.4 | 1085 |
| Claude Fable 5 | 1051 |
| o3 | 967 |
| GPT-5 | 964 |
| Opus 4.8 | 960 |
| GPT-5.2 | 949 |
| GPT-5.4 Nano | 936 |
| GPT-5.6 Terra | 925 |
| GPT-6 Luna | 918 |
| Grok 4.6 | 899 |
| Claude Haiku 5.5 | 874 |
| Gemini 3 Flash Preview | 862 |
| Claude Sonnet 5.5 | 858 |
| Claude Opus 5 | 825 |
| GPT-5.4 Mini | 823 |
| GPT-5 Nano | 809 |
| GPT-5 Mini | 765 |
| Sonnet 5 | 747 |
| Opus 4.6 | 740 |
| Grok 4.20 | 716 |
| Gemini 3.1 Flash-Lite | 705 |
| Grok 4.3 | 688 |