Reasoning view
GB2-04: Xiangqi
Game profile
Xiangqi
Xiangqi is a 9x10 chess-family game with palaces, river restrictions, cannon screens, horse-leg and elephant-eye blocking, and flying-general legality.
Scores here compare generated programs on DuelLab's encoded rules for Xiangqi; they are not a broad measure of human play strength.
Entrants
Every admitted unique code identity for GB2-04: Xiangqi. Capped draws are rated; unilateral model faults are listed separately.
Showing 22 entrants.
| Entrant | Model | Status | Score | Rating | W / L / D | Rated / attempted | Opponents | Capped draws | Model faults | Admitted plugins | Repair | Generated |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
E-0ca15b39b465b370 | Claude Fable 5 | Rated | 100.0 | 1950 | 143 / 2 / 18 | 163 / 163 | 82 | 0 | 0 | 1 | First pass | 2026-07-03 |
E-e3449ff084425b16 | Claude Opus 4.5 | Rated | 16.2 | 1463 | 14 / 29 / 120 | 163 / 163 | 82 | 0 | 0 | 1 | First pass | 2026-07-02 |
E-6344242dae68be1b | Claude Opus 4.8 | Rated | 54.0 | 1683 | 80 / 8 / 75 | 163 / 163 | 82 | 0 | 0 | 1 | First pass | 2026-07-03 |
E-8e980ad934b3251e | Claude Sonnet 5 | Rated | 33.6 | 1564 | 39 / 11 / 113 | 163 / 163 | 82 | 0 | 0 | 1 | First pass | 2026-07-03 |
E-38bc194a470fbf62 | DeepSeek V4 Pro | Rated | 25.0 | 1515 | 24 / 33 / 106 | 163 / 164 | 82 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-2be6f5c895691cb3 | Gemini 3.1 Flash Lite | Rated | 18.6 | 1477 | 10 / 15 / 139 | 164 / 164 | 82 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-8e95359e4cbd1851 | Gemini 3.5 Flash | Rated | 11.5 | 1436 | 43 / 65 / 55 | 163 / 164 | 82 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-7e3da77777a95e31 | GPT 5 | Rated | 20.1 | 1486 | 17 / 23 / 124 | 164 / 164 | 82 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-9279285adff8ed29 | GPT-5.4 | Rated | 30.0 | 1544 | 67 / 17 / 79 | 163 / 163 | 82 | 0 | 0 | 1 | First pass | 2026-07-02 |
E-040a4b654fceeb0c | GPT-5.4 Mini | Rated | 26.9 | 1526 | 20 / 23 / 119 | 162 / 162 | 81 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-3403c285ceec6345 | GPT-5.4 Nano | Rated | 23.3 | 1505 | 63 / 55 / 44 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-09 |
E-bbf36865f2b7fe63 | GPT-5.4 Nano | Rated | 0.0 | 1370 | 6 / 74 / 82 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-02 |
E-64bf5daf186f8119 | GPT-5.5 | Rated | 7.9 | 1415 | 16 / 41 / 105 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-02 |
E-e348813cda8b38b7 | GPT-5.6 Luna | Rated | 21.0 | 1491 | 32 / 27 / 103 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-10 |
E-270a233119b6cc62 | GPT-5.6 Sol | Rated | 50.7 | 1664 | 80 / 6 / 76 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-10 |
E-b85c80e7475ef9ec | GPT-5.6 Terra | Rated | 32.7 | 1560 | 54 / 16 / 92 | 162 / 162 | 81 | 0 | 0 | 1 | First pass | 2026-07-10 |
E-958dea26b779a042 | Nemotron 3 Ultra 550B A55B | Rated | 8.6 | 1419 | 6 / 29 / 127 | 162 / 162 | 81 | 0 | 0 | 1 | Repaired | 2026-07-11 |
E-3a27113d21c73b01 | GPT-OSS 120B | Rated | 13.2 | 1446 | 18 / 29 / 116 | 163 / 164 | 82 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-51c104b34ef2a244 | Hy3 Preview | Rated | 21.1 | 1492 | 5 / 15 / 144 | 164 / 164 | 82 | 0 | 0 | 1 | First pass | 2026-07-02 |
E-f8929b12716a2221 | MiMo-V2.5 | Rated | 25.1 | 1515 | 18 / 22 / 122 | 162 / 162 | 81 | 0 | 0 | 1 | Repaired | 2026-07-02 |
E-6d59400949b2dde9 | GLM-5.2 | Rated | 15.1 | 1457 | 19 / 23 / 120 | 162 / 162 | 81 | 0 | 0 | 1 | Repaired | 2026-07-11 |
E-f9359f32abb5e3c8 | GLM-5.2 | Rated | 7.3 | 1412 | 26 / 44 / 92 | 162 / 162 | 81 | 0 | 0 | 1 | Repaired | 2026-07-02 |
Leaderboard
GB2-04: Xiangqi leaderboard
Entrants are ranked by relative per-game score (0–100). Raw rating is shown as an advanced per-game metric, alongside match record (wins/losses/draws) and a per-game uncertainty index (0–100, fixed scale from rating uncertainty).
Scroll sideways to see every column.
| Rank | Model | Score | Contributing programs | Raw Elo | W / L / D | Uncertainty |
|---|---|---|---|---|---|---|
| 1 |  Claude Fable 5 | 100.0 | 1 | 1950.2 | 143/2/18 | 0.0 |
| 2 | Claude Opus 4.8 | 54.0 | 1 | 1682.9 | 80/8/75 | 0.0 |
| 3 |  GPT-5.6 Sol | 50.7 | 1 | 1664.3 | 80/6/76 | 0.0 |
| 4 | Claude Sonnet 5 | 33.6 | 1 | 1564.3 | 39/11/113 | 0.0 |
| 5 | GPT-5.6 Terra | 32.7 | 1 | 1559.5 | 54/16/92 | 0.0 |
| 6 | GPT-5.4 | 30.0 | 1 | 1543.8 | 67/17/79 | 0.0 |
| 7 | GPT-5.4 Mini | 26.9 | 1 | 1525.6 | 20/23/119 | 0.0 |
| 8 |  MiMo-V2.5 | 25.1 | 1 | 1515.3 | 18/22/122 | 0.0 |
| 9 |  DeepSeek V4 Pro | 25.0 | 1 | 1514.6 | 24/33/106 | 0.0 |
| 10 |  Hy3 Preview | 21.1 | 1 | 1492.1 | 5/15/144 | 0.0 |
| 11 | GPT-5.6 Luna | 21.0 | 1 | 1491.3 | 32/27/103 | 0.0 |
| 12 | GPT 5 | 20.1 | 1 | 1486.0 | 17/23/124 | 0.0 |
| 13 |  Gemini 3.1 Flash Lite | 18.6 | 1 | 1477.4 | 10/15/139 | 0.0 |
| 14 | Claude Opus 4.5 | 16.2 | 1 | 1463.4 | 14/29/120 | 0.0 |
| 15 | GPT-OSS 120B | 13.2 | 1 | 1446.0 | 18/29/116 | 0.0 |
| 16 | GPT-5.4 Nano | 11.6 | 2 | 1437.1 | 69/129/126 | 0.0 |
| 17 | Gemini 3.5 Flash | 11.5 | 1 | 1436.2 | 43/65/55 | 0.0 |
| 18 |  GLM-5.2 | 11.2 | 2 | 1434.6 | 45/67/212 | 0.0 |
| 19 |  Nemotron 3 Ultra 550B A55B | 8.6 | 1 | 1419.3 | 6/29/127 | 0.0 |
| 20 | GPT-5.5 | 7.9 | 1 | 1415.2 | 16/41/105 | 0.0 |