OpenRouter · Vercel · CC BY 4.0
Arena · CC BY 4.0
People compare two answers to the same prompt without knowing which model wrote them, then vote. Only each model’s best-scoring setting is listed.
| Rank | Model | Score | Output price |
|---|---|---|---|
| 1 | Claude Opus 4.6 Anthropic | 1505 pts | $25 per 1M tokens |
| 2 | Claude Fable 5 Anthropic | 1505 pts | $50 per 1M tokens |
| 3 | Claude Opus 5.5 Anthropic | 1504 pts | $20 per 1M tokens |
| 4 | Claude Opus 4.7 Anthropic | 1502 pts | $25 per 1M tokens |
| 5 | Claude Fable 5.1 Anthropic | 1501 pts | $50 per 1M tokens |
| 6 | Muse Spark 1.3 Meta | 1495 pts | $4.25 per 1M tokens |
| 7 | Muse Spark 1.2 Meta | 1494 pts | $4.25 per 1M tokens |
| 8 | Gemini 3.8 Flash Google | 1494 pts | $3.75 per 1M tokens |
| 9 | Muse Spark 1.1 Meta | 1492 pts | $4.25 per 1M tokens |
| 10 | Claude Opus 5 Anthropic | 1491 pts | $25 per 1M tokens |
| 11 | Gemini 3.7 Flash Google | 1488 pts | $3.75 per 1M tokens |
| 12 | Kimi K3 Moonshot | 1488 pts | $13.5 per 1M tokens |
| 13 | Gemini 3.1 Pro Google | 1487 pts | $12 per 1M tokens |
| 14 | GPT-5.6 Sol OpenAI | 1484 pts | $10 per 1M tokens |
| 15 | Gemini 3.6 Flash Google | 1483 pts | $3.75 per 1M tokens |
| 16 | GPT-5.5 OpenAI | 1481 pts | $30 per 1M tokens |
| 17 | Claude Opus 4.8 Anthropic | 1481 pts | $25 per 1M tokens |
| 18 | Qwen3.8 Max Alibaba | 1481 pts | $6 per 1M tokens |
| 19 | MiMo-V2.6-Pro Xiaomi | 1480 pts | $0.87 per 1M tokens |
| 20 | GLM 5.3 Z.ai | 1479 pts | $4.4 per 1M tokens |
| 21 | Gemini 3.5 Flash Google | 1477 pts | $9 per 1M tokens |
| 22 | GPT-6 Astra OpenAI | 1476 pts | $50 per 1M tokens |
Source: Arena text leaderboard (lmarena-ai/leaderboard-dataset), released Sep 30, 2026 · CC BY 4.0 / Prices: OpenRouter model list · Rank changes (▲▼) compare with the previous release