Project AURIONAI Use-Cases Research & Innovation
LMArena (Arena)

Arena Text leaderboard

Crowd-sourced pairwise preference on text chat (arena.ai, formerly LMArena)

human preference arenaactive

Score in Elo, Arena Text overall

  1. 1Google DeepMindGemini 4 Argon1,525 Eloindependent
  2. 2AnthropicClaude Opus 5.51,507 Eloindependent
  3. 3AnthropicClaude Fable 51,504 Eloindependent
36scores
36models
35independent
0maintainer-run
1vendor-reported

Definition

Version
not in the notes
Category
human preference arena
Maintainer
LMArena (Arena)
Status
active
Home page
https://arena.ai/leaderboard/chat/text

Sources

  1. Text Arena: LLM Rankings & AI Model Benchmarks arena.ai/leaderboard/chat/text https://arena.ai/leaderboard/chat/text
  2. Text Arena: LLM Rankings & AI Model Benchmarks arena.ai/leaderboard/text https://arena.ai/leaderboard/text
  3. Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preference arxiv.org https://arxiv.org/abs/2403.04132
  4. The Leaderboard Illusion arxiv.org https://arxiv.org/abs/2504.20879
  5. Arena Leaderboard: Official AI Model Rankings & Benchmarks arena.ai https://arena.ai/leaderboard

Scores at a Glance

Score in Elo, Arena Text overall

Higher is better for this unit. Elo has no zero, so the axis starts at 1,440.

1Google DeepMindGemini 4 ArgonGoogle DeepMindhigh
1,525 Elo
independent2AnthropicClaude Opus 5.5Anthropichigh
1,507 Elo
independent3AnthropicClaude Fable 5Anthropichigh
1,504 Elo
independent4AnthropicClaude Opus 4.6Anthropichigh
1,504 Elo
independent5AnthropicClaude Fable 5.1Anthropicmax
1,501 Elo
independent6Google DeepMindGemini 3.8 FlashGoogle DeepMindhigh
1,497 Elo
independent7MetaMuse Spark 1.3Metamax
1,494 Elo
independent8Moonshot AIKimi K3Moonshot AIkimi-k3-max; top open model
1,488 Elo
independent9Google DeepMindGemini 3.1 Pro (preview)Google DeepMind, Googlepreview
1,487 Elo
independent10OpenAIGPT-5.6 SolOpenAIxhigh
1,485 Elo
independent11OpenAIGPT-6.1 SolOpenAImax
1,484 Elo
independent12Alibaba (Qwen)Qwen3.8 MaxAlibaba (Qwen)
1,483 Elo
independent13XiaomiMiMo-V2.6-ProXiaomi
1,480 Elo
independent14Z.ai (Zhipu)GLM-5.3Z.ai (Zhipu)glm-5.3-max
1,478 Elo
independent15AnthropicClaude Sonnet 5.5Anthropicxhigh
1,476 Elo
independent16DeepSeekDeepSeek-V4.1-FlashDeepSeekdeepseek-v4.1-flash-max
1,475 Elo
independent17OpenAIGPT-6 AstraOpenAImax
1,475 Elo
independent18SpaceXAI (xAI)Grok 4.20SpaceXAI (xAI)grok-4.20-beta1
1,475 Elo
independent19Z.ai (Zhipu)GLM-5.2Z.ai (Zhipu)glm-5.2-max
1,475 Elo
independent20Z.ai (Zhipu)GLM-5.3-FlashZ.ai (Zhipu)
1,475 Elo
independent21XiaomiMiMo-V2.5-ProXiaomi
1,468 Elo
independent22DeepSeekDeepSeek-V4-Pro-0813DeepSeekdeepseek-v4-pro-high-20260813
1,465 Elo
independent23Z.ai (Zhipu)GLM-5.1Z.ai (Zhipu)
1,464 Elo
independent24Moonshot AIKimi K2.6Moonshot AI
1,461 Elo
independent
1,4401,4651,4901,515
Where this comes from
Each bar is one model, with its most independent recorded score. Hatched bars are vendor-reported; the word beside each bar says who ran it. Settings and dates differ between rows; the Scores table below has every result.

Scores

36 of 36 scores shown lmarena-text-scores.csv
ModelScoreWho ran itSettingView or populationDateSource
Google DeepMindGemini 4 Argon1,525 Elo ±9independenthighrank 1, Arena Text overall8 Oct 2026arena.ai
AnthropicClaude Opus 5.51,507 Elo ±8independenthighrank 2, Arena Text overall8 Oct 2026arena.ai
AnthropicClaude Fable 51,504 Elo ±4independenthighrank 4, Arena Text overall8 Oct 2026arena.ai
AnthropicClaude Opus 4.61,504 Elo ±3independenthighrank 3, Arena Text overall8 Oct 2026arena.ai
AnthropicClaude Fable 5.11,501 Elo ±6independentmaxrank 6, Arena Text overall8 Oct 2026arena.ai
Google DeepMindGemini 3.8 Flash1,497 Elo ±5independenthighrank 8, Arena Text overall8 Oct 2026arena.ai
MetaMuse Spark 1.31,494 Elo ±6independentmaxrank 9, Arena Text overall8 Oct 2026arena.ai
Moonshot AIKimi K31,488 Elo ±5independentkimi-k3-max; top open modelrank 16, Arena Text overall8 Oct 2026arena.ai
Google DeepMindGemini 3.1 Pro (preview)1,487 Elo ±3independentpreviewrank 17, Arena Text overall8 Oct 2026arena.ai
OpenAIGPT-5.6 Sol1,485 Elo ±4independentxhighrank 20, Arena Text overall8 Oct 2026arena.ai
OpenAIGPT-6.1 Sol1,484 Elo ±9independentmaxrank 21, Arena Text overall8 Oct 2026arena.ai
Alibaba (Qwen)Qwen3.8 Max1,483 Elo ±5independent–rank 22, Arena Text overall8 Oct 2026arena.ai
XiaomiMiMo-V2.6-Pro1,480 Elo ±9independent–rank 26, Arena Text overall8 Oct 2026arena.ai
Z.ai (Zhipu)GLM-5.31,478 Elo ±5independentglm-5.3-maxrank 27, Arena Text overall8 Oct 2026arena.ai
AnthropicClaude Sonnet 5.51,476 Elo ±9independentxhighrank 32, Arena Text overall8 Oct 2026arena.ai
DeepSeekDeepSeek-V4.1-Flash1,475 Elo ±7independentdeepseek-v4.1-flash-maxrank 39, Arena Text overall8 Oct 2026arena.ai
OpenAIGPT-6 Astra1,475 Elo ±7independentmaxrank 35, Arena Text overall8 Oct 2026arena.ai
SpaceXAI (xAI)Grok 4.201,475 Elo ±5independentgrok-4.20-beta1rank 37, Arena Text overall8 Oct 2026arena.ai
Z.ai (Zhipu)GLM-5.21,475 Elo ±4independentglm-5.2-maxArena Text overall8 Oct 2026arena.ai
Z.ai (Zhipu)GLM-5.3-Flash1,475 Elo ±5independent–rank 38, Arena Text overall8 Oct 2026arena.ai
XiaomiMiMo-V2.5-Pro1,468 Elo ±4independent–Arena Text overall8 Oct 2026arena.ai
DeepSeekDeepSeek-V4-Pro-08131,465 Elo ±7independentdeepseek-v4-pro-high-20260813Arena Text overall8 Oct 2026arena.ai
Z.ai (Zhipu)GLM-5.11,464 Elo ±4independent–Arena Text overall8 Oct 2026arena.ai
Moonshot AIKimi K2.61,461 Elo ±4independent–Arena Text overall8 Oct 2026arena.ai
NVIDIALlama-3.1-Nemotron-70B-Instruct1,267 Elo +-7vendor-reportedChatbot Arena score quoted by NVIDIA on the model card; style controlled rank 26rank 9, Chatbot Arena leaderboard as of 24 Oct 202424 Oct 2024huggingface.co
01.AIyi-lightning– Eloindependent–rank 245, Arena Text overall8 Oct 2026arena.ai
Alibaba (Qwen)qwen3.7-plus– Eloindependent–rank 66, Arena Text overall8 Oct 2026arena.ai
BaiduERNIE 5.1– Eloindependent–rank 49, Arena Text overall8 Oct 2026arena.ai
ByteDanceDola-Seed 2.0 Pro– Eloindependent–rank 65, Arena Text overall8 Oct 2026arena.ai
Coherecommand-a-03-2025– Eloindependent–rank 210, Arena Text overall8 Oct 2026arena.ai
MiniMaxMiniMax-M3– Eloindependent–rank 96, Arena Text overall8 Oct 2026arena.ai
Mistral AIMistral Large 4 (v26.10)– Eloindependent–rank 115, Arena Text overall8 Oct 2026arena.ai
Mistral AIMistral Medium 3.5 (v26.04)– Eloindependent–rank 117, Arena Text overall8 Oct 2026arena.ai
StepFunStep 5 Preview– Eloindependent–rank 83, Arena Text overall8 Oct 2026arena.ai
TencentHy3– Eloindependent–rank 73, Arena Text overall8 Oct 2026arena.ai
UpstageSolar Pro 4– Eloindependentsolar-pro4rank 182, Arena Text overall8 Oct 2026arena.ai
Where this comes from
Scores from different settings, harnesses or dates are not directly comparable; the row says which. Sorted by score, highest first, as a reading aid, not a ranking.

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Arena Text Leaderboard", https://aurion.semai.id/benchmarks/lmarena-text/, records of October 2026.