Historical line
DuelLab v1 compact archive
Compact GameBench 1 rows are frozen historical public records. They use public labels only and should not be compared directly with GameBench 2.
| Rank | Model | Score | Last measured | Notes |
|---|---|---|---|---|
| 1 | anthropic/claude-4.7-opus | 74.4 | 2026-05-13 | Historical public v1 row |
| 2 | gpt-5.4 | 74.2 | 2026-05-13 | Historical public v1 row |
| 3 | moonshotai/kimi-k2.6 | 69.9 | 2026-05-13 | Historical public v1 row |
| 4 | gpt-5.5 | 66.5 | 2026-05-13 | Historical public v1 row |
| 5 | gpt-5.2 | 66.1 | 2026-05-13 | Historical public v1 row |
| 6 | anthropic/claude-opus-4.6 | 65.2 | 2026-05-13 | Historical public v1 row |
| 7 | deepseek/deepseek-v4-pro | 65.0 | 2026-05-13 | Historical public v1 row |
| 8 | z-ai/glm-5.1 | 63.1 | 2026-05-13 | Historical public v1 row |
| 9 | gpt-5.3-codex | 62.4 | 2026-05-13 | Historical public v1 row |
| 10 | google/gemini-3.1-pro-preview | 62.3 | 2026-05-10 | Historical public v1 row |
| 11 | anthropic/claude-sonnet-4.6 | 60.9 | 2026-05-13 | Historical public v1 row |
| 12 | qwen/qwen3.6-plus | 58.6 | 2026-05-10 | Historical public v1 row |
| 13 | moonshotai/kimi-k2.5 | 58.2 | 2026-05-13 | Historical public v1 row |
| 14 | gpt-5.4-nano | 55.6 | 2026-05-13 | Historical public v1 row |
| 15 | z-ai/glm-5 | 55.5 | 2026-05-13 | Historical public v1 row |
| 16 | xiaomi/mimo-v2.5-pro | 54.2 | 2026-05-13 | Historical public v1 row |
| 17 | openrouter/owl-alpha | 53.3 | 2026-06-21 | Historical public v1 row |
| 18 | deepseek/deepseek-v4-flash | 52.5 | 2026-05-13 | Historical public v1 row |
| 19 | gpt-5.2-codex | 51.1 | 2026-05-13 | Historical public v1 row |
| 20 | xiaomi/mimo-v2-pro | 50.5 | 2026-05-13 | Historical public v1 row |
Compatibility
Read as history, not a v2 rank
Do not compare GameBench 1 ranks directly with GameBench 2 ranks. The game sets and evaluation contexts are different.