Human preference · Elo
LMArena Text
Crowd-sourced arena (formerly LMSYS Chatbot Arena): users chat with two anonymous models and vote for the better answer; votes are turned into a Bradley–Terry rating.
Active Contamination risk: Low 414 results · 381 models Updated Oct 22, 2026 Newest result: 2026-10-08 Official leaderboard ↗
What it measures
Human preference in open-ended chat (overall category, with style control).
Limitations
Preference is not correctness; answer length and formatting sway votes; providers can test many private variants before release; scores are relative and drift as models are added.
Contamination risk: Low
Fresh user prompts, but providers may tune for arena style.
Leaderboard
Top 100 of 381
| # | Model | Provider | Score | Rank | Setting | Source type | Date |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 4 Argon | 1525 | 1 | Overall, style control on; 4892 votes; 95% CI 1517–1534“gemini-4-argon-high” | Official leaderboard | Oct 8, 2026 | |
| 2 | Claude Opus 5.5 | Anthropic | 1507 | 2 | Overall, style control on; 6272 votes; 95% CI 1499–1515“claude-opus-5.5-high” | Official leaderboard | Oct 8, 2026 |
| 3 | Claude Opus 4.6 | Anthropic | 1504 | 3 | Overall, style control on; 79415 votes; 95% CI 1501–1508“claude-opus-4-6-high” | Official leaderboard | Oct 8, 2026 |
| 4 | Claude Fable 5 | Anthropic | 1504 | 4 | Overall, style control on; 40026 votes; 95% CI 1499–1508“claude-fable-5-high” | Official leaderboard | Oct 8, 2026 |
| 5 | Claude Opus 4.7 | Anthropic | 1501 | 5 | Overall, style control on; 65192 votes; 95% CI 1497–1505“claude-opus-4-7-high” | Official leaderboard | Oct 8, 2026 |
| 6 | claude-fable-5.1-max | Anthropic | 1501 | 6 | Overall, style control on; 13063 votes; 95% CI 1494–1507 | Official leaderboard | Oct 8, 2026 |
| 7 | Gemini 3.8 Flash | 1497 | 8 | Overall, style control on; 31437 votes; 95% CI 1492–1501“gemini-3.8-flash-high” | Official leaderboard | Oct 8, 2026 | |
| 8 | muse-spark-1.3-max | Meta | 1494 | 9 | Overall, style control on; 14694 votes; 95% CI 1488–1500 | Official leaderboard | Oct 8, 2026 |
| 9 | muse-spark-1.2 (xHigh) | Meta | 1492 | 11 | Overall, style control on; 6118 votes; 95% CI 1484–1500 | Official leaderboard | Oct 8, 2026 |
| 10 | Muse Spark 1.1 | Meta | 1491 | 12 | Overall, style control on; 39847 votes; 95% CI 1486–1495“muse-spark-1.1” | Official leaderboard | Oct 8, 2026 |
| 11 | Claude Opus 5 | Anthropic | 1490 | 13 | Overall, style control on; 62934 votes; 95% CI 1486–1494“claude-opus-5-high” | Official leaderboard | Oct 8, 2026 |
| 12 | claude-opus-5-max | Anthropic | 1489 | 14 | Overall, style control on; 30569 votes; 95% CI 1485–1494 | Official leaderboard | Oct 8, 2026 |
| 13 | Muse Spark | Meta | 1489 | 15 | Overall, style control on; 14168 votes; 95% CI 1483–1495“muse-spark” | Official leaderboard | Oct 8, 2026 |
| 14 | kimi-k3-maxOW | Moonshot AI | 1488 | 16 | Overall, style control on; 30351 votes; 95% CI 1483–1493 | Official leaderboard | Oct 8, 2026 |
| 15 | Gemini 3.1 Pro | 1487 | 17 | Overall, style control on; 124675 votes; 95% CI 1484–1490“gemini-3.1-pro-preview” | Official leaderboard | Oct 8, 2026 | |
| 16 | Gemini 3.7 Flash | 1487 | 18 | Overall, style control on; 24173 votes; 95% CI 1481–1492“gemini-3.7-flash-high” | Official leaderboard | Oct 8, 2026 | |
| 17 | Gemini 3 Pro | 1486 | 19 | Overall, style control on; 42078 votes; 95% CI 1482–1489“gemini-3-pro” | Official leaderboard | Oct 8, 2026 | |
| 18 | GPT-5.6 Sol | OpenAI | 1485 | 20 | Overall, style control on; 38490 votes; 95% CI 1481–1489“gpt-5.6-sol-xhigh” | Official leaderboard | Oct 8, 2026 |
| 19 | gpt-6.1-sol-max | OpenAI | 1484 | 21 | Overall, style control on; 4512 votes; 95% CI 1475–1493 | Official leaderboard | Oct 8, 2026 |
| 20 | Qwen 3.8 Max | Alibaba (Qwen) | 1483 | 22 | Overall, style control on; 25378 votes; 95% CI 1478–1488“qwen3.8-max” | Official leaderboard | Oct 8, 2026 |
| 21 | Gemini 3.6 Flash | 1482 | 23 | Overall, style control on; 36172 votes; 95% CI 1478–1487“gemini-3.6-flash-high” | Official leaderboard | Oct 8, 2026 | |
| 22 | GPT-5.5 | OpenAI | 1482 | 24 | Overall, style control on; 69437 votes; 95% CI 1478–1485“gpt-5.5-high” | Official leaderboard | Oct 8, 2026 |
| 23 | Claude Opus 4.8 | Anthropic | 1481 | 25 | Overall, style control on; 64177 votes; 95% CI 1478–1485“claude-opus-4-8-high” | Official leaderboard | Oct 8, 2026 |
| 24 | MiMo-V2.6 ProOW | Xiaomi Corp | 1480 | 26 | Overall, style control on; 4069 votes; 95% CI 1471–1490“mimo-v2.6-pro” | Official leaderboard | Oct 8, 2026 |
| 25 | glm-5.3-maxOW | Zhipu AI (Z.ai) | 1478 | 27 | Overall, style control on; 20049 votes; 95% CI 1473–1484 | Official leaderboard | Oct 8, 2026 |
| 26 | Gemini 3.5 Flash | 1478 | 28 | Overall, style control on; 48533 votes; 95% CI 1474–1481“gemini-3.5-flash-high” | Official leaderboard | Oct 8, 2026 | |
| 27 | gpt-5.2-chat-latest | OpenAI | 1476 | 31 | Overall, style control on; 36007 votes; 95% CI 1472–1480“gpt-5.2-chat-latest-20260210” | Official leaderboard | Oct 8, 2026 |
| 28 | Claude Sonnet 5.5 | Anthropic | 1476 | 32 | Overall, style control on; 4838 votes; 95% CI 1467–1485“claude-sonnet-5.5-xhigh” | Official leaderboard | Oct 8, 2026 |
| 29 | glm-5.2-maxOW | Zhipu AI (Z.ai) | 1475 | 33 | Overall, style control on; 47521 votes; 95% CI 1471–1479 | Official leaderboard | Oct 8, 2026 |
| 30 | GPT-5.4 | OpenAI | 1475 | 34 | Overall, style control on; 64844 votes; 95% CI 1471–1479“gpt-5.4-high” | Official leaderboard | Oct 8, 2026 |
| 31 | gpt-6-astra-max | OpenAI | 1475 | 35 | Overall, style control on; 10536 votes; 95% CI 1469–1482 | Official leaderboard | Oct 8, 2026 |
| 32 | qwen3.7-max-preview | Alibaba (Qwen) | 1475 | 36 | Overall, style control on; 3932 votes; 95% CI 1465–1485 | Official leaderboard | Oct 8, 2026 |
| 33 | grok-4.20-beta1 | xAI | 1475 | 37 | Overall, style control on; 27928 votes; 95% CI 1470–1479 | Official leaderboard | Oct 8, 2026 |
| 34 | GLM-5.3-FlashOW | Zhipu AI (Z.ai) | 1475 | 38 | Overall, style control on; 27678 votes; 95% CI 1470–1480“glm-5.3-flash” | Official leaderboard | Oct 8, 2026 |
| 35 | deepseek-v4.1-flash-maxOW | DeepSeek | 1475 | 39 | Overall, style control on; 10872 votes; 95% CI 1468–1481 | Official leaderboard | Oct 8, 2026 |
| 36 | Claude Opus 4.5 | Anthropic | 1474 | 41 | Overall, style control on; 37527 votes; 95% CI 1470–1478“claude-opus-4-5-20251101-high-32k” | Official leaderboard | Oct 8, 2026 |
| 37 | GPT-5.5 Instant | OpenAI | 1473 | 42 | Overall, style control on; 27187 votes; 95% CI 1468–1478“gpt-5.5-instant” | Official leaderboard | Oct 8, 2026 |
| 38 | Gemini 3 Flash | 1473 | 43 | Overall, style control on; 31337 votes; 95% CI 1469–1477“gemini-3-flash” | Official leaderboard | Oct 8, 2026 | |
| 39 | Claude Sonnet 4.6 | Anthropic | 1472 | 44 | Overall, style control on; 70971 votes; 95% CI 1468–1476“claude-sonnet-4-6” | Official leaderboard | Oct 8, 2026 |
| 40 | grok-4.20-beta-0309-reasoning | xAI | 1472 | 45 | Overall, style control on; 66524 votes; 95% CI 1468–1475 | Official leaderboard | Oct 8, 2026 |
| 41 | grok-4.20-multi-agent-beta-0309 | xAI | 1471 | 46 | Overall, style control on; 65008 votes; 95% CI 1467–1474 | Official leaderboard | Oct 8, 2026 |
| 42 | MiMo-V2.5-ProOW | Xiaomi Corp | 1468 | 48 | Overall, style control on; 73197 votes; 95% CI 1464–1471“mimo-v2.5-pro” | Official leaderboard | Oct 8, 2026 |
| 43 | ernie-5.1 | Baidu | 1467 | 49 | Overall, style control on; 39536 votes; 95% CI 1463–1472 | Official leaderboard | Oct 8, 2026 |
| 44 | GPT-5.6 Terra | OpenAI | 1466 | 50 | Overall, style control on; 40278 votes; 95% CI 1462–1470“gpt-5.6-terra-xhigh” | Official leaderboard | Oct 8, 2026 |
| 45 | Grok 4.5 | xAI | 1466 | 51 | Overall, style control on; 39953 votes; 95% CI 1461–1470“grok-4.5” | Official leaderboard | Oct 8, 2026 |
| 46 | grok-4.1-thinking | xAI | 1465 | 52 | Overall, style control on; 66511 votes; 95% CI 1462–1468 | Official leaderboard | Oct 8, 2026 |
| 47 | qwen3.5-max-preview | Alibaba (Qwen) | 1465 | 54 | Overall, style control on; 22624 votes; 95% CI 1460–1470 | Official leaderboard | Oct 8, 2026 |
| 48 | DeepSeek-V4-ProOW | DeepSeek | 1465 | 55 | Overall, style control on; 9815 votes; 95% CI 1458–1471“deepseek-v4-pro-high-20260813” | Official leaderboard | Oct 8, 2026 |
| 49 | GLM-5.1OW | Zhipu AI (Z.ai) | 1464 | 56 | Overall, style control on; 60592 votes; 95% CI 1461–1468“glm-5.1” | Official leaderboard | Oct 8, 2026 |
| 50 | Claude Sonnet 5 | Anthropic | 1461 | 57 | Overall, style control on; 47576 votes; 95% CI 1457–1465“claude-sonnet-5-high” | Official leaderboard | Oct 8, 2026 |
| 51 | Kimi K2.6OW | Moonshot AI | 1461 | 58 | Overall, style control on; 40066 votes; 95% CI 1456–1465“kimi-k2.6” | Official leaderboard | Oct 8, 2026 |
| 52 | Qwen 3.6 Max (Preview) | Alibaba (Qwen) | 1460 | 59 | Overall, style control on; 5426 votes; 95% CI 1452–1469“qwen3.6-max-preview” | Official leaderboard | Oct 8, 2026 |
| 53 | GLM-5OW | Zhipu AI (Z.ai) | 1458 | 63 | Overall, style control on; 29292 votes; 95% CI 1453–1462“glm-5” | Official leaderboard | Oct 8, 2026 |
| 54 | Claude Sonnet 4.5 | Anthropic | 1456 | 64 | Overall, style control on; 83973 votes; 95% CI 1454–1459“claude-sonnet-4-5-20250929-high-32k” | Official leaderboard | Oct 8, 2026 |
| 55 | dola-seed-2.0-pro | ByteDance | 1456 | 65 | Overall, style control on; 79195 votes; 95% CI 1453–1460 | Official leaderboard | Oct 8, 2026 |
| 56 | Qwen3.7-Plus | Alibaba (Qwen) | 1456 | 66 | Overall, style control on; 42073 votes; 95% CI 1452–1460“qwen3.7-plus” | Official leaderboard | Oct 8, 2026 |
| 57 | gpt-6-sol-max | OpenAI | 1456 | 67 | Overall, style control on; 10469 votes; 95% CI 1449–1463 | Official leaderboard | Oct 8, 2026 |
| 58 | GPT-5.1 | OpenAI | 1456 | 68 | Overall, style control on; 41674 votes; 95% CI 1452–1459“gpt-5.1-high” | Official leaderboard | Oct 8, 2026 |
| 59 | Gemini 3.5 Flash-Lite | 1455 | 70 | Overall, style control on; 35892 votes; 95% CI 1451–1460“gemini-3.5-flash-lite” | Official leaderboard | Oct 8, 2026 | |
| 60 | deepseek-v4-pro-high-previewOW | DeepSeek | 1455 | 71 | Overall, style control on; 54976 votes; 95% CI 1451–1459 | Official leaderboard | Oct 8, 2026 |
| 61 | Grok 4.6 | xAI | 1454 | 72 | Overall, style control on; 26115 votes; 95% CI 1449–1459“grok-4.6-high” | Official leaderboard | Oct 8, 2026 |
| 62 | hy3OW | Tencent | 1454 | 73 | Overall, style control on; 11284 votes; 95% CI 1448–1461 | Official leaderboard | Oct 8, 2026 |
| 63 | GPT-5.6 Luna | OpenAI | 1452 | 74 | Overall, style control on; 40917 votes; 95% CI 1448–1457“gpt-5.6-luna-xhigh” | Official leaderboard | Oct 8, 2026 |
| 64 | gemma-4-31bOW | 1452 | 75 | Overall, style control on; 6145 votes; 95% CI 1445–1460 | Official leaderboard | Oct 8, 2026 | |
| 65 | MiMo-V2.6 FlashOW | Xiaomi Corp | 1451 | 76 | Overall, style control on; 8226 votes; 95% CI 1444–1458“mimo-v2.6-flash” | Official leaderboard | Oct 8, 2026 |
| 66 | gpt-5.3-chat-latest | OpenAI | 1450 | 77 | Overall, style control on; 34367 votes; 95% CI 1446–1454 | Official leaderboard | Oct 8, 2026 |
| 67 | Kimi K2.5OW | Moonshot AI | 1450 | 78 | Overall, style control on; 74811 votes; 95% CI 1447–1453“kimi-k2.5-thinking” | Official leaderboard | Oct 8, 2026 |
| 68 | Claude Opus 4.1 | Anthropic | 1450 | 79 | Overall, style control on; 49555 votes; 95% CI 1446–1453“claude-opus-4-1-20250805-thinking-16k” | Official leaderboard | Oct 8, 2026 |
| 69 | ernie-5.0-preview-1203 | Baidu | 1449 | 80 | Overall, style control on; 9816 votes; 95% CI 1442–1455 | Official leaderboard | Oct 8, 2026 |
| 70 | MiMo-V2-Pro | Xiaomi Corp | 1448 | 82 | Overall, style control on; 25447 votes; 95% CI 1443–1452“mimo-v2-pro” | Official leaderboard | Oct 8, 2026 |
| 71 | Step 5 Preview | StepFun | 1447 | 83 | Overall, style control on; 4799 votes; 95% CI 1439–1456 | Official leaderboard | Oct 8, 2026 |
| 72 | GPT-5.4 Mini | OpenAI | 1447 | 84 | Overall, style control on; 63493 votes; 95% CI 1443–1451“gpt-5.4-mini-high” | Official leaderboard | Oct 8, 2026 |
| 73 | ernie-5.0-0110 | Baidu | 1446 | 85 | Overall, style control on; 36525 votes; 95% CI 1442–1450 | Official leaderboard | Oct 8, 2026 |
| 74 | Gemini 2.5 Pro (Jun 2025) | 1446 | 86 | Overall, style control on; 125379 votes; 95% CI 1443–1448“gemini-2.5-pro” | Official leaderboard | Oct 8, 2026 | |
| 75 | GPT-4.5 | OpenAI | 1445 | 87 | Overall, style control on; 14547 votes; 95% CI 1439–1450“gpt-4.5-preview-2025-02-27” | Official leaderboard | Oct 8, 2026 |
| 76 | Qwen 3.6 Plus | Alibaba (Qwen) | 1443 | 88 | Overall, style control on; 48208 votes; 95% CI 1439–1447“qwen3.6-plus” | Official leaderboard | Oct 8, 2026 |
| 77 | chatgpt-4o-latest | OpenAI | 1443 | 89 | Overall, style control on; 81927 votes; 95% CI 1440–1446“chatgpt-4o-latest-20250326” | Official leaderboard | Oct 8, 2026 |
| 78 | Grok 4.7 | xAI | 1443 | 90 | Overall, style control on; 8780 votes; 95% CI 1436–1450“grok-4.7-xhigh” | Official leaderboard | Oct 8, 2026 |
| 79 | gpt-6-luna-max | OpenAI | 1443 | 91 | Overall, style control on; 28632 votes; 95% CI 1438–1448 | Official leaderboard | Oct 8, 2026 |
| 80 | Qwen3.5 397B-A17BOW | Alibaba (Qwen) | 1442 | 92 | Overall, style control on; 91190 votes; 95% CI 1439–1445“qwen3.5-397b-a17b” | Official leaderboard | Oct 8, 2026 |
| 81 | Grok 4.3 Beta | xAI | 1442 | 93 | Overall, style control on; 71448 votes; 95% CI 1438–1445“grok-4.3” | Official leaderboard | Oct 8, 2026 |
| 82 | GLM-4.7OW | Zhipu AI (Z.ai) | 1441 | 94 | Overall, style control on; 12256 votes; 95% CI 1435–1447“glm-4.7” | Official leaderboard | Oct 8, 2026 |
| 83 | InklingOW | Thinking Machines | 1441 | 95 | Overall, style control on; 37744 votes; 95% CI 1437–1445“inkling” | Official leaderboard | Oct 8, 2026 |
| 84 | MiniMax-M3OW | MiniMax | 1441 | 96 | Overall, style control on; 61220 votes; 95% CI 1437–1444“minimax-m3” | Official leaderboard | Oct 8, 2026 |
| 85 | deepseek-v4-flash-high-previewOW | DeepSeek | 1439 | 97 | Overall, style control on; 51859 votes; 95% CI 1435–1443 | Official leaderboard | Oct 8, 2026 |
| 86 | Qwen 3.8 27BOW | Alibaba (Qwen) | 1438 | 99 | Overall, style control on; 20711 votes; 95% CI 1433–1443“qwen3.8-27b” | Official leaderboard | Oct 8, 2026 |
| 87 | Gemma 4 26B A4BOW | 1438 | 100 | Overall, style control on; 6092 votes; 95% CI 1430–1445“gemma-4-26b-a4b” | Official leaderboard | Oct 8, 2026 | |
| 88 | GPT-5.2 | OpenAI | 1437 | 101 | Overall, style control on; 49685 votes; 95% CI 1434–1441“gpt-5.2-high” | Official leaderboard | Oct 8, 2026 |
| 89 | longcat-flash-chat-2602-exp | Meituan | 1436 | 102 | Overall, style control on; 29306 votes; 95% CI 1432–1441 | Official leaderboard | Oct 8, 2026 |
| 90 | deepseek-v4-flashOW | DeepSeek | 1436 | 103 | Overall, style control on; 52047 votes; 95% CI 1432–1440 | Official leaderboard | Oct 8, 2026 |
| 91 | GPT-5 | OpenAI | 1435 | 105 | Overall, style control on; 31723 votes; 95% CI 1430–1439“gpt-5-high” | Official leaderboard | Oct 8, 2026 |
| 92 | qwen3-max-preview | Alibaba (Qwen) | 1434 | 106 | Overall, style control on; 27379 votes; 95% CI 1430–1439 | Official leaderboard | Oct 8, 2026 |
| 93 | MiMo-V2.5OW | Xiaomi Corp | 1434 | 107 | Overall, style control on; 47402 votes; 95% CI 1429–1438“mimo-v2.5” | Official leaderboard | Oct 8, 2026 |
| 94 | glm-5v-turbo | Zhipu AI (Z.ai) | 1434 | 108 | Overall, style control on; 10492 votes; 95% CI 1427–1440 | Official leaderboard | Oct 8, 2026 |
| 95 | Gemini-3.1-Flash-Lite-Preview | 1433 | 109 | Overall, style control on; 64257 votes; 95% CI 1429–1437“gemini-3.1-flash-lite-preview” | Official leaderboard | Oct 8, 2026 | |
| 96 | o3 | OpenAI | 1432 | 110 | Overall, style control on; 58913 votes; 95% CI 1428–1435“o3-2025-04-16” | Official leaderboard | Oct 8, 2026 |
| 97 | mimo-v2-omni | Xiaomi | 1431 | 111 | Overall, style control on; 20667 votes; 95% CI 1425–1436 | Official leaderboard | Oct 8, 2026 |
| 98 | Grok 4.1 fast | xAI | 1430 | 112 | Overall, style control on; 57574 votes; 95% CI 1427–1434“grok-4-1-fast-reasoning” | Official leaderboard | Oct 8, 2026 |
| 99 | Kimi K2 ThinkingOW | Moonshot AI | 1430 | 114 | Overall, style control on; 63390 votes; 95% CI 1427–1433“kimi-k2-thinking-turbo” | Official leaderboard | Oct 8, 2026 |
| 100 | Mistral Large 4 | Mistral AI | 1429 | 115 | Overall, style control on; 4179 votes; 95% CI 1420–1439“mistral-large-4” | Official leaderboard | Oct 8, 2026 |
Data sources & licences
| Source | Licence / terms | How we reuse it | Last fetched |
|---|---|---|---|
| LMArena leaderboard dataset LMArena, Arena Leaderboard Dataset (Hugging Face), CC BY 4.0 | CC BY 4.0 | Open licence – full data with attribution | Oct 10, 2026 ⚠︎ |
FAQ
What does LMArena Text measure?
Crowd-sourced arena (formerly LMSYS Chatbot Arena): users chat with two anonymous models and vote for the better answer; votes are turned into a Bradley–Terry rating. Human preference in open-ended chat (overall category, with style control).
How reliable is LMArena Text?
Preference is not correctness; answer length and formatting sway votes; providers can test many private variants before release; scores are relative and drift as models are added. Contamination risk: Low – Fresh user prompts, but providers may tune for arena style.