Skip to content
IFAI

Model rankings

892 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Text arena · Anonymous head-to-head votes · Published December 23, 2025

Most preferred
gemini-3-proGoogle1,490Score
Best open model
kimi-k2-thinking-turboMoonshot AI1,428Score
Best from China
ernie-5.0-preview-1203Baidu1,446Score
Most voted
claude-3-opus-20240229Anthropic196,365Votes
Top three by Score
1gemini-3-proGoogle1,490
2gemini-3-flashGoogle1,478
Top three open models
1kimi-k2-thinking-turboMoonshot AI · #22 overall1,428
2glm-4.6Z.ai · #24 overall1,424
3deepseek-v3.2-expDeepSeek · #26 overall1,423
This week's changes
glm-4.7Coding · #5New
MiMo-V2-Flash (Reasoning)Intelligence · #10New
K2-V2 (high)Intelligence · #67New
RankChangeModelScore95% intervalVotesPrice $/M
1—ernie-5.0-preview-1203ClosedBaidu1,446±113,185—
2—qwen3-max-previewClosedAlibaba1,433±528,023—
3—kimi-k2-thinking-turboOpenMoonshot AI1,428±519,526—
4—glm-4.6OpenZ.ai1,424±427,979—
5—qwen3-max-2025-09-23ClosedAlibaba1,424±69,222—
6—deepseek-v3.2-expOpenDeepSeek · 1 more variant1,423±611,926—
7—deepseek-v3.2-thinkingOpenDeepSeek · 1 more variant1,422±79,657—
8—qwen3-235b-a22b-instruct-2507OpenAlibaba1,421±455,241—
9—kimi-k2-0905-previewOpenMoonshot AI · 1 more variant1,419±711,873—
10—deepseek-r1-0528OpenDeepSeek · 1 more variant1,418±619,162—
11—deepseek-v3.1OpenDeepSeek · 1 more variant1,417±615,210—
12—deepseek-v3.1-terminusOpenDeepSeek · 1 more variant1,415±103,733—
13—qwen3-vl-235b-a22b-instructOpenAlibaba1,414±79,449—
14—glm-4.5OpenZ.ai1,409±524,723—
15—qwen3-next-80b-a3b-instructOpenAlibaba1,400±523,005—
16—longcat-flash-chatOpenMeituan1,400±611,465—
17—qwen3-235b-a22b-no-thinkingOpenAlibaba1,400±439,231—
18—qwen3-235b-a22b-thinking-2507OpenAlibaba1,397±69,310—
19—qwen3-vl-235b-a22b-thinkingOpenAlibaba1,394±77,959—
20—deepseek-v3-0324OpenDeepSeek · 1 more variant1,392±446,613—
21—hunyuan-vision-1.5-thinkingClosedTencent1,391±122,204—
22Newmimo-v2-flash (non-thinking)OpenXiaomi1,388±122,604—
23▼ 1qwen3-coder-480b-a35b-instructOpenAlibaba1,386±524,124—
24▼ 1hunyuan-t1-20250711ClosedTencent1,385±94,802—
25▼ 1qwen3-30b-a3b-instruct-2507OpenAlibaba1,382±524,105—
26▼ 1hunyuan-turbos-20250416ClosedTencent · 1 more variant1,380±611,094—
27▼ 1qwen3-235b-a22bOpenAlibaba1,374±527,066—
28▼ 1qwen2.5-maxClosedAlibaba1,373±433,497—
29▼ 1glm-4.5-airOpenZ.ai1,370±431,522—
30▼ 1qwen3-next-80b-a3b-thinkingOpenAlibaba1,367±613,762—
31▼ 1minimax-m1OpenMiniMax1,366±436,709—
32▼ 1glm-4.5vOpenZ.ai1,352±84,967—
33—step-3OpenStepFun1,346±76,611—
34▼ 2ling-flash-2.0OpenAnt Group1,346±77,120—
35▼ 1minimax-m2OpenMiniMax1,346±87,084—
36▼ 1qwen3-32bOpenAlibaba1,345±93,944—
37▼ 1qwen-plus-0125ClosedAlibaba1,344±85,862—
38▼ 1glm-4-plus-0111ClosedZhipu · 1 more variant1,343±85,808—
39▼ 1hunyuan-turbo-0110ClosedTencent1,339±112,324—
40▼ 1qwq-32bOpenAlibaba · 1 more variant1,334±426,194—
41▼ 1step-2-16k-exp-202412ClosedStepFun1,332±94,896—
42▼ 1yi-lightningClosed01 AI1,328±527,624—
43▼ 1qwen3-30b-a3bOpenAlibaba1,326±527,398—
44▼ 1hunyuan-large-2025-02-10ClosedTencent1,325±103,760—
45▼ 1deepseek-v2.5-1210OpenDeepSeek · 1 more variant1,322±86,877—
46▼ 1ring-flash-2.0OpenAnt Group1,320±77,246—
47▼ 1step-1o-turbo-202506ClosedStepFun1,320±79,635—
48▼ 1qwen-max-0919OpenAlibaba1,317±616,599—
49▼ 1qwen2.5-plus-1127ClosedAlibaba1,314±610,254—
50▼ 1hunyuan-standard-2025-02-10ClosedTencent1,310±103,923—

Intelligence over time

Each maker's best model at the time on the Artificial Analysis Intelligence Index, by the month it was released.

Gained most · Anthropic +57.2Highest now · Anthropic 64.6Gap between highest and lowest 18.1 ptsTime range Nov 2025 – Nov 2026
64.6Anthropic60.0Google56.9OpenAI50.1Meta46.5Kimi

Arena data comes from the leaderboard dataset it publishes on Hugging Face (CC BY 4.0). Only each board's overall leaderboard is used; scores and ranks are unchanged. dataset · CC BY 4.0

Artificial Analysis data comes from its API (artificialanalysis.ai). Scores are unchanged; ranks follow the scores. artificialanalysis.ai

Open or closed follows the license Arena gives; prices and Artificial Analysis intelligence show only where model names match, otherwise “—”.