Skip to content
IFAI

Model rankings

855 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Text arena · Anonymous head-to-head votes · Published December 5, 2025

Most preferred
gemini-3-proGoogle1,491Score
Best open model
kimi-k2-thinking-turboMoonshot AI1,428Score
Best from China
qwen3-max-previewAlibaba1,433Score
Most voted
claude-3-opus-20240229Anthropic196,368Votes
Top three by Score
1gemini-3-proGoogle1,491
Top three open models
1kimi-k2-thinking-turboMoonshot AI · #17 overall1,428
2glm-4.6Z.ai · #19 overall1,426
3deepseek-v3.2-expDeepSeek · #23 overall1,423
This week's changes
Runway Gen-4.5Video · #1New
Claude Opus 4.5 (Reasoning)Intelligence · #2New
FLUX.2 [pro]Image · #2New
Riverflow 2 PreviewImage · #4New
RankChangeModelScore95% intervalVotesPrice $/M
1—qwen3-max-previewClosedAlibaba1,433±528,132—
2—kimi-k2-thinking-turboOpenMoonshot AI1,428±613,647—
3—glm-4.6OpenZ.ai1,426±522,680—
4▲ 1qwen3-max-2025-09-23ClosedAlibaba1,424±69,256—
5▼ 1deepseek-v3.2-expOpenDeepSeek · 1 more variant1,423±711,969—
6—qwen3-235b-a22b-instruct-2507OpenAlibaba1,421±449,337—
7Newdeepseek-v3.2-thinkingOpenDeepSeek · 1 more variant1,418±103,983—
8▼ 1kimi-k2-0905-previewOpenMoonshot AI · 1 more variant1,418±711,841—
9—deepseek-r1-0528OpenDeepSeek · 1 more variant1,418±619,240—
10▼ 2ernie-5.0-preview-1022ClosedBaidu1,417±94,704—
11▼ 1deepseek-v3.1OpenDeepSeek · 1 more variant1,417±615,255—
12—deepseek-v3.1-terminusOpenDeepSeek · 1 more variant1,415±103,746—
13▼ 2qwen3-vl-235b-a22b-instructOpenAlibaba1,415±78,535—
14▼ 1glm-4.5OpenZ.ai1,409±524,825—
15▼ 1qwen3-next-80b-a3b-instructOpenAlibaba1,400±523,118—
16▼ 1longcat-flash-chatOpenMeituan1,400±611,502—
17▼ 1qwen3-235b-a22b-no-thinkingOpenAlibaba1,399±439,377—
18▼ 1qwen3-235b-a22b-thinking-2507OpenAlibaba1,397±69,346—
19▼ 1qwen3-vl-235b-a22b-thinkingOpenAlibaba1,394±77,986—
20▼ 1deepseek-v3-0324OpenDeepSeek · 1 more variant1,392±446,784—
21▼ 1hunyuan-vision-1.5-thinkingClosedTencent1,391±122,212—
22▼ 1qwen3-coder-480b-a35b-instructOpenAlibaba1,385±523,154—
23▼ 1hunyuan-t1-20250711ClosedTencent1,385±94,819—
24▼ 1qwen3-30b-a3b-instruct-2507OpenAlibaba1,382±524,196—
25▼ 1hunyuan-turbos-20250416ClosedTencent · 1 more variant1,380±611,130—
26▼ 1qwen3-235b-a22bOpenAlibaba1,374±527,168—
27▼ 1qwen2.5-maxClosedAlibaba1,373±433,548—
28▼ 1glm-4.5-airOpenZ.ai1,370±431,673—
29▼ 1qwen3-next-80b-a3b-thinkingOpenAlibaba1,367±613,822—
30▼ 1minimax-m1OpenMiniMax1,366±436,884—
31▼ 1glm-4.5vOpenZ.ai1,352±84,978—
32▼ 1ling-flash-2.0OpenAnt Group1,346±77,158—
33▼ 1minimax-m2OpenMiniMax1,346±87,123—
34▼ 1step-3OpenStepFun1,346±76,642—
35▼ 1qwen3-32bOpenAlibaba1,345±93,943—
36▼ 1qwen-plus-0125ClosedAlibaba1,345±85,861—
37▼ 1glm-4-plus-0111ClosedZhipu · 1 more variant1,343±85,806—
38▼ 1hunyuan-turbo-0110ClosedTencent1,339±112,322—
39▼ 1qwq-32bOpenAlibaba · 1 more variant1,334±426,273—
40▼ 1step-2-16k-exp-202412ClosedStepFun1,332±94,895—
41▼ 1yi-lightningClosed01 AI1,328±527,624—
42▼ 1qwen3-30b-a3bOpenAlibaba1,326±527,492—
43▼ 1hunyuan-large-2025-02-10ClosedTencent1,325±103,760—
44▼ 1deepseek-v2.5-1210OpenDeepSeek · 1 more variant1,322±86,877—
45▼ 1ring-flash-2.0OpenAnt Group1,320±77,282—
46▼ 1step-1o-turbo-202506ClosedStepFun1,320±79,665—
47▼ 1qwen-max-0919OpenAlibaba1,317±616,598—
48▼ 1qwen2.5-plus-1127ClosedAlibaba1,314±610,252—
49▼ 1hunyuan-standard-2025-02-10ClosedTencent1,310±103,920—
50▼ 1qwen2.5-72b-instructOpenAlibaba1,302±439,632—

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 “—”.