Skip to content
IFAI

Model rankings

451 models · 2 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

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

Most preferred
gemini-2.5-proGoogle1,458Score
Best open model
Best from China
Most voted
claude-3-opus-20240229Anthropic202,641Votes
Top three by Score
1gemini-2.5-proGoogle1,458
2o3-2025-04-16OpenAI1,451
Top three open models
1qwen3-235b-a22b-instruct-2507Alibaba · #5 overall1,433
2kimi-k2-0711-previewMoonshot AI · #7 overall1,420
3deepseek-r1-0528DeepSeek · #9 overall1,417
This week's changes
Qwen3 Coder 480B A35BIntelligence · #32New
Granite 3.3 Instruct 8BIntelligence · #130New
Gemma 3n E2B InstructIntelligence · #144New
Qwen3 235B A22B 2507 (Reasoning)Intelligence · #6Up 8
RankChangeModelScore95% intervalVotesPrice $/M
1—gemini-2.5-proClosedGoogle1,458±525,480—
2—o3-2025-04-16ClosedOpenAI1,451±531,450—
3—chatgpt-4o-latest-20250326ClosedOpenAI1,442±530,344—
4—gpt-4.5-preview-2025-02-27ClosedOpenAI1,438±615,271—
5—grok-4-0709ClosedxAI1,429±612,591—
6—claude-opus-4-20250514-thinking-16kClosedAnthropic1,420±617,512—
7—claude-opus-4-20250514ClosedAnthropic1,412±525,729—
8—grok-3-preview-02-24ClosedxAI1,410±531,348—
9—gemini-2.5-flashClosedGoogle1,409±530,765—
10—gpt-4.1-2025-04-14ClosedOpenAI1,408±524,381—
11—claude-sonnet-4-20250514-thinking-32kClosedAnthropic1,400±616,700—
12—o1-2024-12-17ClosedOpenAI1,398±429,038—
13—o4-mini-2025-04-16ClosedOpenAI1,396±524,037—
14—claude-sonnet-4-20250514ClosedAnthropic1,388±522,210—
15▲ 1claude-3-7-sonnet-20250219-thinking-32kClosedAnthropic1,386±532,207—
16▼ 1o1-previewClosedOpenAI1,385±533,177—
17—gpt-4.1-mini-2025-04-14ClosedOpenAI1,371±523,459—
18—claude-3-7-sonnet-20250219ClosedAnthropic1,369±437,199—
19▲ 1claude-3-5-sonnet-20241022ClosedAnthropic · 1 more variant1,365±384,101—
20▼ 1gemini-2.0-flash-001ClosedGoogle1,365±440,716—
21▲ 1gemma-3-27b-itOpenGoogle1,363±431,799—
22▼ 1o3-mini-highClosedOpenAI1,363±519,404—
23—grok-3-mini-highClosedxAI1,359±79,780—
24—grok-3-mini-betaClosedxAI1,354±616,456—
25—gemini-2.0-flash-lite-preview-02-05ClosedGoogle1,351±426,104—
26—gemini-1.5-pro-002ClosedGoogle1,350±358,645—
27—llama-3.1-nemotron-ultra-253b-v1OpenNVIDIA1,344±122,656—
28—o3-miniClosedOpenAI1,344±442,792—
29—gpt-4o-2024-05-13ClosedOpenAI1,343±3117,747—
30—gemma-3-12b-itOpenGoogle1,340±93,976—
31▲ 1o1-miniClosedOpenAI1,335±454,951—
32▼ 1llama-3.1-405b-instruct-bf16OpenMeta1,334±443,788—
33—llama-3.1-405b-instruct-fp8OpenMeta1,332±363,038—
34—gpt-4o-2024-08-06ClosedOpenAI1,332±447,973—
35—grok-2-2024-08-13ClosedxAI1,331±367,084—
36—gemini-advanced-0514ClosedGoogle1,331±552,144—
37—llama-4-maverick-17b-128e-instructOpenMeta1,330±524,088—
38—llama-3.3-nemotron-49b-super-v1OpenNVIDIA1,324±122,371—
39—gpt-4-turbo-2024-04-09ClosedOpenAI1,323±4102,133—
40—claude-3-opus-20240229ClosedAnthropic1,321±3202,641—
41—gemini-1.5-pro-001ClosedGoogle1,320±482,435—
42—gpt-4.1-nano-2025-04-14ClosedOpenAI1,319±86,302—
43▲ 1llama-3.3-70b-instructOpenMeta1,316±451,606—
44▲ 1claude-3-5-haiku-20241022ClosedAnthropic1,316±355,498—
45▲ 1gpt-4o-mini-2024-07-18ClosedOpenAI1,315±372,429—
46▼ 3llama-4-scout-17b-16e-instructOpenMeta1,315±613,727—
47▲ 2gemma-3n-e4b-itOpenGoogle1,314±612,786—
48▼ 1gpt-4-1106-previewClosedOpenAI1,314±4103,748—
49▼ 1gpt-4-0125-previewClosedOpenAI1,314±497,079—
50—gemini-1.5-flash-002ClosedGoogle1,311±437,021—

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 “—”.