Who’s Winning AI? Watch the live race

LMArena text scores · as of 19 October 2026

AI model leaderboard

Gemini 4 Argon from Google DeepMind is the highest-rated AI model, with an LMArena text score of 1524. Here are the 87 most important models from the 9 labs in the AI Race, ranked.

#ModelLabScoreReleased
1Gemini 4 Argon Google DeepMind15242026-10
2Muse Wonder Meta15092026-10
3Claude Opus 5.5 Anthropic15082026-09
4Claude Fable 5 Anthropic15042026-06
5Claude Opus 4.6 Anthropic15032026-02
6Gemini 3.8 Flash Google DeepMind14962026-09
7Muse Spark 1.2 Meta14932026-08
8Muse Spark 1.1 Meta14902026-07
9Muse Spark Meta14892026-04
10Gemini 3.7 Flash Google DeepMind14882026-08
11Gemini 3.1 Pro Preview Google DeepMind14872026-02
12Kimi K3 Moonshot AI14872026-07
13GPT-5.6 Sol OpenAI14852026-07
14Gemini 3 Pro Google DeepMind14852025-11
15Qwen3.8-Max Alibaba Qwen14822026-08
16GPT-5.5 OpenAI14812026-04
17GPT-5.2 OpenAI14762026-02
18Grok 4.20 Beta 1 xAI14752026-02
19DeepSeek-V4.1-Flash DeepSeek14752026-09
20Claude Opus 4.5 Anthropic14742025-11
21Qwen3.7-Max-Preview Alibaba Qwen14742026-05
22Grok 4.1 xAI14652025-11
23DeepSeek-V4-Pro DeepSeek14652026-08
24Qwen3.6-Max-Preview Alibaba Qwen14612026-04
25Kimi K2.6 Moonshot AI14612026-04
26DeepSeek-V4-Pro DeepSeek14582026-04
27GPT-5.1 OpenAI14562025-11
28Claude Sonnet 4.5 Anthropic14562025-09
29DeepSeek-V4-Pro Preview DeepSeek14552026-04
30Claude Opus 4.1 Anthropic14502025-08
31Kimi K2.5 Moonshot AI14502026-01
32GPT-4.5 Preview OpenAI14452025-02
33Gemini 2.5 Pro Google DeepMind14452025-03
34Qwen3.6-Plus Alibaba Qwen14432026-04
35Qwen3.5-397B-A17B Alibaba Qwen14422026-02
36Qwen3-Max-Preview Alibaba Qwen14342025-09
37Mistral Large 4 Mistral AI14322026-10
38Kimi K2 Thinking Turbo Moonshot AI14302025-11
39Claude Opus 4 Anthropic14262025-05
40Mistral Medium 3.5 Mistral AI14262026-04
41DeepSeek-V3.2-Exp DeepSeek14252025-09
42Qwen3-235B-A22B Alibaba Qwen14222025-07
43DeepSeek-R1 DeepSeek14212025-05
44Grok 4 Fast xAI14192025-09
45Kimi K2 Preview Moonshot AI14192025-09
46Kimi K2 Preview Moonshot AI14182025-07
47Mistral Large 3 Mistral AI14142025-12
48Grok 3 Preview xAI14112025-02
49Mistral Medium 3.1 Mistral AI14082025-08
50o1 OpenAI14022024-12
51Qwen3-235B-A22B Alibaba Qwen14022025-04
52DeepSeek-R1 DeepSeek13982025-01
53o1-preview OpenAI13892024-09
54Claude 3.7 Sonnet Anthropic13882025-02
55Mistral Medium 3 Mistral AI13872025-05
56Claude 3.5 Sonnet Anthropic13742024-10
57Qwen2.5-Max Alibaba Qwen13742025-01
58Gemini 2.0 Flash Google DeepMind13602025-02
59DeepSeek-V3 DeepSeek13582024-12
60Gemini 1.5 Pro Google DeepMind13512024-09
61GPT-4o OpenAI13462024-05
62Claude 3.5 Sonnet Anthropic13432024-06
63Gemini Advanced Google DeepMind13362024-05
64Grok 2 xAI13362024-08
65Llama 3.1 405B Meta13352024-07
66GPT-4 Turbo OpenAI13252024-04
67Gemini 1.5 Pro Google DeepMind13242024-04
68DeepSeek-V2.5 DeepSeek13232024-12
69Claude 3 Opus Anthropic13222024-03
70Qwen-Max Alibaba Qwen13182024-09
71Mistral Large 2 Mistral AI13142024-07
72GPT-4 Turbo Preview OpenAI13132023-11
73DeepSeek-V2.5 DeepSeek13072024-09
74GPT-4 OpenAI12882023-03
75Llama 3 70B Meta12772024-04
76DeepSeek-Coder-V2 DeepSeek12652024-06
77Qwen2-72B Alibaba Qwen12622024-06
78Mistral Large Mistral AI12432024-02
79Gemini Pro Google DeepMind12372023-12
80Qwen1.5-72B Alibaba Qwen12342024-02
81Mistral Medium Mistral AI12232023-12
82Mixtral 8x7B Mistral AI11972023-12
83DeepSeek LLM 67B DeepSeek11852023-11
84Llama 2 70B Meta11712023-07
85PaLM 2 Google DeepMind11392023-05
86Qwen-14B Alibaba Qwen11392023-09
87Mistral 7B Mistral AI11102023-09

Scores are LMArena’s text leaderboard (default view), where people compare two anonymous models and vote for the better answer. Differences under about 5 points are within the margin of error. How AI benchmarks work.

Head-to-head comparisons

Watch the live AI Race