Skip to content
IFAI

Model rankings

420 models · 2 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Text arena · Anonymous head-to-head votes · Published July 1, 2025

Most preferred
gemini-2.5-proGoogle1,463Score
Best open model
deepseek-r1-0528DeepSeek1,413Score
Best from China
deepseek-r1-0528DeepSeek1,413Score
Most voted
claude-3-opus-20240229Anthropic202,641Votes
Top three by Score
1gemini-2.5-proGoogle1,463
2o3-2025-04-16OpenAI1,449
Top three open models
1deepseek-r1-0528DeepSeek · #7 overall1,413
2qwen3-235b-a22b-no-thinkingAlibaba · #12 overall1,396
3deepseek-v3-0324DeepSeek · #13 overall1,396
This week's changes
Solar Pro 2 (Preview) (Reasoning)Intelligence · #38New
Gemma 3n E4B InstructIntelligence · #110New
Gemma 2 27BIntelligence · #102Up 20
Gemma 3 4B InstructIntelligence · #124Up 2
RankChangeModelScore95% intervalVotesPrice $/M
1—gemini-2.5-proClosedGoogle1,463±514,062—
2—o3-2025-04-16ClosedOpenAI1,449±420,095—
3—chatgpt-4o-latest-20250326ClosedOpenAI1,441±423,599—
4—gpt-4.5-preview-2025-02-27ClosedOpenAI1,436±515,271—
5—claude-opus-4-20250514ClosedAnthropic1,417±419,708—
6—gemini-2.5-flashClosedGoogle1,415±519,451—
7—gpt-4.1-2025-04-14ClosedOpenAI1,411±417,456—
8—grok-3-preview-02-24ClosedxAI1,408±325,067—
9—o1-2024-12-17ClosedOpenAI1,399±329,038—
10—o4-mini-2025-04-16ClosedOpenAI1,398±517,150—
11—claude-sonnet-4-20250514ClosedAnthropic1,393±515,543—
12—claude-3-7-sonnet-20250219-thinking-32kClosedAnthropic1,385±425,145—
13—o1-previewClosedOpenAI1,383±333,177—
14—claude-3-7-sonnet-20250219ClosedAnthropic1,375±429,775—
15—gpt-4.1-mini-2025-04-14ClosedOpenAI1,373±516,454—
16—claude-3-5-sonnet-20241022ClosedAnthropic · 1 more variant1,366±277,078—
17—o3-mini-highClosedOpenAI1,366±419,404—
18—gemini-2.0-flash-001ClosedGoogle1,364±336,673—
19—gemma-3-27b-itOpenGoogle1,363±425,902—
20—grok-3-mini-betaClosedxAI1,356±69,759—
21—gemini-2.0-flash-lite-preview-02-05ClosedGoogle1,350±326,104—
22—gemini-1.5-pro-002ClosedGoogle1,348±258,645—
23—o3-miniClosedOpenAI1,347±436,135—
24—gpt-4o-2024-05-13ClosedOpenAI1,343±2117,747—
25—gemma-3-12b-itOpenGoogle1,342±93,976—
26—o1-miniClosedOpenAI1,335±254,951—
27—gemini-advanced-0514ClosedGoogle1,334±352,144—
28—llama-3.1-405b-instruct-bf16OpenMeta1,334±343,788—
29—llama-3.1-405b-instruct-fp8OpenMeta1,333±263,038—
30▲ 1llama-4-maverick-17b-128e-instructOpenMeta1,330±517,075—
31▼ 1gpt-4o-2024-08-06ClosedOpenAI1,330±347,973—
32—grok-2-2024-08-13ClosedxAI1,329±267,084—
33—llama-3.3-nemotron-49b-super-v1OpenNVIDIA1,326±112,371—
34—gpt-4-turbo-2024-04-09ClosedOpenAI1,322±2102,133—
35—gemini-1.5-pro-001ClosedGoogle1,321±282,435—
36—gpt-4.1-nano-2025-04-14ClosedOpenAI1,320±76,302—
37—claude-3-opus-20240229ClosedAnthropic1,319±2202,641—
38▲ 5gemma-3n-e4b-itOpenGoogle1,316±76,311—
39—claude-3-5-haiku-20241022ClosedAnthropic1,315±348,556—
40▲ 1llama-3.3-70b-instructOpenMeta1,315±347,256—
41▼ 1gpt-4-1106-previewClosedOpenAI1,315±2103,748—
42—gpt-4o-mini-2024-07-18ClosedOpenAI1,314±272,531—
43▲ 1gpt-4-0125-previewClosedOpenAI1,311±297,079—
44▲ 1gemini-1.5-flash-002ClosedGoogle1,309±337,021—
45▲ 1grok-2-mini-2024-08-13ClosedxAI1,304±355,442—
46▲ 1gemma-3-4b-itOpenGoogle1,301±104,321—
47▲ 1llama-3.1-nemotron-70b-instructOpenNVIDIA1,296±57,577—
48▲ 1llama-3.1-70b-instructOpenMeta1,294±258,637—
49▲ 1llama-3.1-tulu-3-70bOpenAi21,289±93,010—
50▲ 1amazon-nova-pro-v1.0ClosedAmazon1,288±326,371—

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 “—”.