Benchmark · Other
Text Arena
Arena score Higher is better Benchmark homepage
What it measures
Text Arena, run by Arena (arena.ai, formerly LMArena), ranks models from crowdsourced blind pairwise votes on text prompts covering math, coding, creative writing and other open-ended tasks. Each model receives a score with a confidence interval and a vote count. As of 2026-10-08, the Overall leaderboard covered 414 models and 8,734,330 votes. This is a human-preference rating, not a correctness test.
Caveats:
Scores measure voter preference, which can be influenced by answer style and length. The leaderboard offers a style-control adjustment, so record which view was used. Scores shift with every update, so a value is only meaningful together with its leaderboard date. Entries marked "Preliminary" or with few votes have wide confidence intervals. The same model can appear several times with different reasoning-effort suffixes (for example -high or -max). Category views (Coding, Math, Creative Writing and others) are different rankings from Overall.
Results
| Model | Score | Setting | Reported by | Measured | Verified | Source |
|---|---|---|---|---|---|---|
| Gemini 4 Argon Google | 1,517 | Overall, style control on; high effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Muse Wonder Meta | 1,509 | Overall, style control on (default); max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Claude Opus 5.5 Anthropic | 1,507 | Overall, style control on; high effort | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Claude Fable 5.1 Anthropic | 1,501 | Overall, style control on; max effort | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Gemini 3.8 Flash Google | 1,496 | Overall, style control on; high effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Muse Spark 1.3 Meta | 1,496 | Overall, style control on (default); max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Muse Spark 1.2 Meta | 1,491 | Overall, style control on (default); xhigh effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Muse Spark 1.1 Meta | 1,490 | Overall, style control on (default); effort not stated | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Gemini 3.1 Pro Google | 1,487 | Overall, style control on; effort not in model name | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Gemini 3.6 Flash Google | 1,483 | Overall, style control on; high effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Qwen3.8-Max Alibaba Cloud (Qwen) | 1,483 | Overall, style control on; no effort suffix | Third party | 2026-10-22 | 2026-10-25 | Text Arena leaderboard (Arena) |
| GPT-6.1 Sol OpenAI | 1,480 | Overall, style control on; max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Claude Sonnet 5.5 Anthropic | 1,476 | Overall, style control on; xhigh effort | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| DeepSeek-V4.1-Flash DeepSeek | 1,476 | Overall, style control on; max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| GPT-6 Astra OpenAI | 1,476 | Overall, style control on; max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Grok 4.20 (Reasoning) xAI | 1,472 | Overall; style control on; reasoning variant | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Grok 4.20 Multi-Agent (Beta) xAI | 1,471 | Overall; style control on; multi-agent | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Grok 4.5 xAI | 1,466 | Overall; style control on; effort not stated | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| DeepSeek-V4-Pro DeepSeek | 1,465 | Overall, style control on; high effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| GPT-6 Sol OpenAI | 1,459 | Overall, style control on; max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Claude Sonnet 4.5 Anthropic | 1,456 | Overall, style control on; thinking, high-32k | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Grok 4.6 xAI | 1,456 | Overall; style control on; high effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Qwen3.7-Plus Alibaba Cloud (Qwen) | 1,456 | Overall, style control on; no effort suffix | Third party | 2026-10-22 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Claude Sonnet 4.5 Anthropic | 1,455 | Overall, style control on; no effort suffix | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Gemini 3.5 Flash-Lite Google | 1,455 | Overall, style control on; effort not in model name | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Gemma 4 31B Google | 1,452 | Overall, style control on; effort not in model name | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| GPT-6 Luna OpenAI | 1,442 | Overall, style control on; max effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Grok 4.3 xAI | 1,442 | Overall; style control on; effort not stated | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Grok 4.7 xAI | 1,441 | Overall; style control on; xhigh effort | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Qwen3.8-27B Alibaba Cloud (Qwen) | 1,437 | Overall, style control on; no effort suffix | Third party | 2026-10-22 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Mistral Large 4 Mistral AI | 1,434 | Overall, style control (default view); no effort suffix | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Mistral Medium 3.5 Mistral AI | 1,426 | Overall, style control (default view); no effort suffix | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Muse Glimmer Meta | 1,425 | Overall, style control on (default); effort not stated | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Mistral Large 3 Mistral AI | 1,418 | LMArena text leaderboard, view and style control not stated; non-reasoning model | Self-reported | 2025-12-02 | 2026-10-25 | Introducing Mistral 3 | Mistral AI |
| Mistral Large 3 Mistral AI | 1,414 | Overall, style control (default view); no effort suffix | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Command A Cohere | 1,354 | Overall; style control on (default view); no effort suffix (row 'command-a-03-2025') | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| gpt-oss-120b OpenAI | 1,352 | Overall, style control on | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Amazon Nova 2 Lite Amazon | 1,335 | Overall; style control on (default view); effort not stated | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Llama 4 Maverick Meta | 1,327 | Overall, style control on (default) | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Llama 4 Scout Meta | 1,322 | Overall, style control on (default) | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| gpt-oss-20b OpenAI | 1,317 | Overall, style control on | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Amazon Nova Pro Amazon | 1,290 | Overall; style control on (default view) | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Amazon Nova Lite Amazon | 1,260 | Overall; style control on (default view) | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
| Command R (08-2024) Cohere | 1,250 | Overall; style control on (default view); no effort suffix (row 'command-r-08-2024') | Third party | 2026-10-23 | 2026-10-25 | Text Arena leaderboard (Arena) |
| Amazon Nova Micro Amazon | 1,241 | Overall; style control on (default view) | Third party | 2026-10-08 | 2026-10-12 | Text Arena leaderboard (Arena) |
Sorted by score, without rank numbers: each result was measured under its own setting (effort level, harness, tools, test version), as stated by the source. Scores are only directly comparable between rows with the same setting. Self-reported results come from the provider and have not been reproduced independently.