Benchmarks · reasoning
Lichess Puzzle Elo (Dubesor)
LLMs solve rated Lichess puzzles (100+ of a 500-puzzle set); the value is a maximum-likelihood performance rating against the puzzles' Lichess ratings, first-try solves only.
47 models · 48 results · updated 2026-10-31 · dubesor.de ↗
Leaderboard
Gemini 4 Argon high
More from this benchmark
3 published measuresScore vs any published measure3 measures: sub-scores, splits, counts
Breakdown
darker is better within each column · click a column to sort| Model | ||
|---|---|---|
| Gemini 4 Argon high | 74.0 | 250.0 |
| Gemini 3.7 Flash max | 54.0 | 500.0 |
| Gemini 3.8 Flash max | 53.0 | 500.0 |
| Gemini 3.1 Pro high | 81.0 | 200.0 |
| Gemini 3.6 Flash xhigh | 52.0 | 500.0 |
| GPT-6 Astra max | 71.0 | 250.0 |
| DeepSeek V4.1 Flash 1017 max | 81.0 | 200.0 |
| GPT-6.1 Sol max | 51.0 | 500.0 |
| Opus 5.5 max, fallback | 72.0 | 250.0 |
| Gemini 3 Flash high | 51.0 | 500.0 |
| Hy3 xhigh | 72.0 | 250.0 |
| DeepSeek V4 Flash 0731 max | 81.0 | 200.0 |
| GPT-6.1 Terra max | 51.0 | 500.0 |
| Muse Spark 1.3 max | 55.0 | 424.0 |
| MiniMax M3.1 max | 81.0 | 200.0 |
| GPT-5.6 Terra max | 81.0 | 200.0 |
| Sonnet 5.5 max, fallback | 81.0 | 200.0 |
| GPT-5.6 Luna max | 51.0 | 500.0 |
| Haiku 5.5 max | 51.0 | 500.0 |
| gpt-oss-120b high | 51.0 | 500.0 |
| GPT-6 Luna Pro max | 84.0 | 190.0 |
| GPT-6 Luna max | 52.0 | 500.0 |
| Opus 4.8 max | 119.0 | 100.0 |
| Gemini 3.1 Flash-Lite high | 52.0 | 500.0 |
| gpt-oss-20b high | 53.0 | 500.0 |
All results
48 rows| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Gemini 4 Argon | high | 2150 ±74 |
| 2 | Gemini 3.7 Flash | max | 2100 ±54 |
| 3 | Gemini 3.8 Flash | max | 2091 ±53 |
| 4 | Gemini 3.1 Pro | high | 2060 ±81 |
| 5 | Gemini 3.6 Flash | xhigh | 1963 ±52 |
| 6 | GPT-6 Astra | max | 1921 ±71 |
| 7 | Gemini 4 Argon | max | 1843 ±51 |
| 8 | DeepSeek V4.1 Flash 1017 | max | 1760 ±81 |
| 9 | GPT-6.1 Sol | max | 1736 ±51 |
| 10 | Opus 5.5 | max | 1631 ±72 |
| 11 | Gemini 3 Flash | high | 1630 ±51 |
| 12 | Hy3 | xhigh | 1616 ±72 |
| 13 | DeepSeek V4 Flash 0731 | max | 1600 ±81 |
| 14 | GPT-6.1 Terra | max | 1571 ±51 |
| 15 | Muse Spark 1.3 | max | 1516 ±55 |
| 16 | MiniMax M3.1 | max | 1503 ±81 |
| 17 | GPT-5.6 Terra | max | 1384 ±81 |
| 18 | Sonnet 5.5 | max | 1325 ±81 |
| 19 | GPT-5.6 Luna | max | 1317 ±51 |
| 20 | Haiku 5.5 | max | 1239 ±51 |
| 21 | gpt-oss-120b | high | 1183 ±51 |
| 22 | GPT-6 Luna Pro | max | 1163 ±84 |
| 23 | GPT-6 Luna | max | 1103 ±52 |
| 24 | Opus 4.8 | max | 1083 ±119 |
| 25 | Gemini 3.1 Flash-Lite | high | 1083 ±52 |
| 26 | gpt-oss-20b | high | 1029 ±53 |
| 27 | Qwen 3.7 Flash | max | 986 ±54 |
| 28 | Gemini 3.5 Flash-Lite | xhigh | 973 ±54 |
| 29 | GPT-4.1 | high | 915 ±55 |
| 30 | Gemma 4 31B IT | high | 863 ±57 |