| # | Model | Arena score | Samples |
|---|---|---|---|
| 1 |  gpt-6-astra-maxOpenAI | 1,800±16 | 2,281 |
| 2 |  claude-fable-5.1-maxAnthropic | 1,758±14 | 3,036 |
| 3 | claude-opus-5-maxAnthropic | 1,687±7 | 12,087 |
| 4 | qwen3.8-max-0902AlibabaEarly | 1,681±15 | 2,262 |
| 5 |  kimi-k3-maxMoonshot | 1,674±11 | 4,547 |
| 6 | qwen3.8-maxAlibabaEarly | 1,671±12 | 3,221 |
| 7 |
| claude-opus-5-highAnthropic |
|---|
| 1,660±7 |
| 12,566 |
| 8 | 1,652±12 | 2,972 |
| 9 | qwen3.8-flash-nextAlibabaEarly | 1,635±13 | 2,779 |
|---|
| 10 | claude-fable-5Anthropic | 1,628±7 | 10,081 |
|---|
| 11 | hy4-previewTencentEarly | 1,624±13 | 2,321 |
|---|
| 12 | 1,623±14 | 2,123 |
| 13 | grok-4.6-highSpaceXAI | 1,618±10 | 4,282 |
|---|
| 14 | 1,617±7 | 11,916 |
| 15 |  glm-5.3-maxZ.ai | 1,614±11 | 3,725 |
|---|
| 16 |  deepseek-v4.1-flash-maxDeepSeek | 1,614±17 | 1,361 |
|---|
| 17 | glm-5.3-flashZ.ai | 1,607±12 | 2,925 |
|---|
| 18 | qwen3.8-27bAlibaba | 1,593±9 | 4,913 |
|---|
| 19 | glm-5.2-maxZ.ai | 1,592±7 | 10,516 |
|---|
| 20 |  gemini-3.7-flash-highGoogleEarly | 1,587±12 | 3,007 |
|---|
Dots: score · lines: published uncertainty