Skip to content
IFAI

Model rankings

420 models · 2 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated July 5, 2025

Smartest
o3-proOpenAI71.0Intelligence
Best open model
DeepSeek R1 0528 (May '25)DeepSeek68.3Intelligence
Best value
Gemma 3n E4B InstructGoogle1,120 pts/$Value
Best from China
DeepSeek R1 0528 (May '25)DeepSeek68.3Intelligence
Top three by Intelligence
1o3-proOpenAI71.0
2Gemini 2.5 ProGoogle70.5
3o3OpenAI70.0
Top three open models
1DeepSeek R1 0528 (May '25)DeepSeek · #6 overall68.3
2MiniMax M1 80kMiniMax · #11 overall63.0
3Qwen3 235B A22B (Reasoning)Alibaba · #12 overall62.3
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
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—o3-proClosedOpenAI71.0$35——
2—Gemini 2.5 ProClosedGoogle70.5$3.4——
3—o3ClosedOpenAI70.0$3.5——
4—o4-mini (high)ClosedOpenAI69.8$1.9——
5—Gemini 2.5 Pro Preview (Mar' 25)ClosedGoogle68.9$3.4——
6—DeepSeek R1 0528 (May '25)OpenDeepSeek68.3$0.96——
7—Gemini 2.5 Pro Preview (May' 25)ClosedGoogle67.9$3.4——
8—Grok 3 mini Reasoning (high)ClosedxAI66.7$0.35——
9—o3-mini (high)ClosedOpenAI65.5$1.9——
10—Gemini 2.5 Flash (Reasoning)ClosedGoogle65.1$0.99——
11—Claude 4 Opus (Extended Thinking)ClosedAnthropic64.4$30——
12—MiniMax M1 80kOpenMiniMax63.0$0.83——
13—o3-miniClosedOpenAI62.9$1.9——
14—Qwen3 235B A22B (Reasoning)OpenAlibaba62.3$2.6——
15—o1ClosedOpenAI61.9$26——
16—MiniMax M1 40kOpenMiniMax61.4$0.83——
17—Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)OpenNVIDIA60.8$0.90——
18—Claude 4 Sonnet (Extended Thinking)ClosedAnthropic60.8$6.0——
19—Gemini 2.5 Flash Preview (Reasoning)ClosedGoogle60.5$0.99——
20—DeepSeek R1 (Jan '25)OpenDeepSeek60.2$2.0——
21—o1-previewClosedOpenAI59.6$26——
22—Qwen3 32B (Reasoning)OpenAlibaba59.2$2.6——
23—QwQ 32BOpenAlibaba58.1$0.33——
24—Claude 4 OpusClosedAnthropic57.7$30——
25—Claude 3.7 Sonnet (Extended Thinking)ClosedAnthropic57.4$6.0——
26—Grok 3 Reasoning BetaClosedxAI56.1$0.000——
26—o1-proClosedOpenAI56.1$263——
28—Magistral MediumClosedMistral56.0$2.8——
29—Qwen3 14B (Reasoning)OpenAlibaba55.7$1.3——
30—Qwen3 30B A3B (Reasoning)OpenAlibaba55.6$0.75——
31—Gemini 2.5 Flash-Lite (Reasoning)ClosedGoogle54.9$0.18——
32—Magistral SmallOpenMistral54.7$0.75——
33—o1-miniClosedOpenAI53.8$1.9——
34—Gemini 2.5 FlashClosedGoogle53.4$0.26——
35—DeepSeek V3 0324 (Mar '25)OpenDeepSeek53.2$0.48——
36—Claude 4 SonnetClosedAnthropic53.0$6.0——
36—GPT-4.5 (Preview)ClosedOpenAI53.0$94——
38—GPT-4.1 miniClosedOpenAI52.9$0.70——
39—GPT-4.1ClosedOpenAI52.6$3.5——
40—Gemini 2.0 Flash Thinking Experimental (Jan '25)ClosedGoogle52.3$0.000——
41—DeepSeek R1 0528 Qwen3 8BOpenDeepSeek52.2$0.065——
42—DeepSeek R1 Distill Qwen 32BOpenDeepSeek51.9$0.30——
43—Qwen3 8B (Reasoning)OpenAlibaba51.4$0.66——
44—Llama 3.3 Nemotron Super 49B v1 (Reasoning)OpenNVIDIA51.2$0.000——
45NewSolar Pro 2 (Preview) (Reasoning)ClosedUpstage50.8$0.000——
46▼ 1Grok 3ClosedxAI50.6$6.0——
47▼ 1Llama 4 MaverickOpenMeta50.5$0.39——
48▼ 1GPT-4o (March 2025, chatgpt-4o-latest)ClosedOpenAI50.3$7.5——
49▼ 1Gemini 2.0 Pro Experimental (Feb '25)ClosedGoogle49.2$0.000——
50▼ 1DeepSeek R1 Distill Qwen 14BOpenDeepSeek49.0$0.20——

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 “—”.