Chatbot Arena
Crowdsourced human preference Elo ratings. All models.
Top 25 models
| Rank | Model | Score | Captured |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 1503 | 2026-09-13 |
| 2 | Claude Opus 4.7 | 1490 | 2026-09-13 |
| 3 | Gemini 3.1 Pro | 1480 | 2026-09-13 |
| 4 | Gemini 3 Pro | 1479 | 2026-09-13 |
| 5 | GLM-5 | 1475 | 2026-09-13 |
| 6 | GPT-5 | 1471 | 2026-09-13 |
| 7 | GPT-5.4 | 1470 | 2026-09-13 |
| 8 | Gemini 3 Flash | 1467 | 2026-09-13 |
| 9 | Claude Opus 4 | 1461 | 2026-09-13 |
| 10 | Claude Sonnet 4.6 | 1458 | 2026-09-13 |
| 11 | Gemini 2.5 Pro | 1458 | 2026-09-13 |
| 12 | Kimi K2 | 1455 | 2026-09-13 |
| 13 | Grok 4 | 1451 | 2026-09-13 |
| 14 | Claude Opus 4.5 | 1450 | 2026-09-13 |
| 15 | GPT-5.1 | 1442 | 2026-09-13 |
| 16 | GLM-4.6 | 1440 | 2026-09-13 |
| 17 | GPT-5.2 | 1439 | 2026-09-13 |
| 18 | Claude Sonnet 4.5 | 1438 | 2026-09-13 |
| 19 | GLM-4.7 | 1436 | 2026-09-13 |
| 20 | GPT-4o | 1429 | 2026-09-13 |
| 21 | DeepSeek R1 | 1428 | 2026-09-13 |
| 22 | Mistral Large | 1427 | 2026-09-13 |
| 23 | Grok 3 | 1425 | 2026-09-13 |
| 24 | DeepSeek V3 | 1425 | 2026-09-13 |
| 25 | DeepSeek V3.1 | 1420 | 2026-09-13 |
Upstream leaderboard: https://lmarena.ai/leaderboard