Skip to content
IFAI

Model rankings

288 models · 2 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated March 26, 2025

Smartest
o3-mini (high)OpenAI65.5Intelligence
Best open model
DeepSeek R1DeepSeek60.2Intelligence
Best from China
DeepSeek R1DeepSeek60.2Intelligence
Top three by Intelligence
1o3-mini (high)OpenAI65.5
2o1OpenAI61.9
3DeepSeek R1DeepSeek60.2
Top three open models
1DeepSeek R1DeepSeek · #3 overall60.2
2QwQ 32BAlibaba · #4 overall58.1
3DeepSeek R1 Distill Qwen 32BDeepSeek · #7 overall51.9
This week's changes
Llama 3.3 Nemotron Super 49B v1Intelligence · #29New
Grok 2 (Dec '24)Intelligence · #30New
Mistral Small 3.1Intelligence · #41New
Gemini 1.5 Flash-8BIntelligence · #50New
Gemma 2 9BIntelligence · #71New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—DeepSeek R1OpenDeepSeek60.2———
2—QwQ 32BOpenAlibaba58.1———
3—DeepSeek R1 Distill Qwen 32BOpenDeepSeek51.9———
4▲ 3DeepSeek V3 (Mar' 25)OpenDeepSeek · 1 more variant51.8———
5▼ 1DeepSeek R1 Distill Qwen 14BOpenDeepSeek49.0———
6▼ 1DeepSeek R1 Distill Llama 70BOpenDeepSeek48.2———
7▼ 1Reka Flash 3OpenReka AI47.1———
8—QwQ 32B-PreviewOpenAlibaba42.6———
9—Llama 3.3 Instruct 70BOpenMeta41.1———
10—Llama 3.1 Instruct 405BOpenMeta40.5———
11—Qwen2.5 Instruct 72BOpenAlibaba40.4———
12▲ 3MiniMax-Text-01OpenMiniMax40.2———
13▼ 1Phi-4OpenMicrosoft Azure40.2———
14▼ 1Command AOpenCohere40.0———
15▼ 1Llama 3.1 Tulu3 405BOpenAllen Institute for AI39.9———
16NewLlama 3.3 Nemotron Super 49B v1OpenNVIDIA39.3———
17▼ 1Mistral Large 2 (Nov '24)OpenMistral · 1 more variant38.3———
18▼ 1Gemma 3 27B InstructOpenGoogle37.6———
19▼ 1Pixtral LargeOpenMistral37.4———
20▼ 1Qwen2.5 Instruct 32BOpenAlibaba37.4———
21▼ 1Llama 3.1 Nemotron Instruct 70BOpenNVIDIA37.3———
22▼ 1Qwen2.5 Coder Instruct 32BOpenAlibaba36.3———
23▼ 1Llama 3.1 Instruct 70BOpenMeta35.4———
24NewMistral Small 3.1OpenMistral35.3———
25▼ 2Mistral Small 3OpenMistral35.3———
26▼ 2DeepSeek R1 Distill Llama 8BOpenDeepSeek34.0———
27▼ 2Llama 3.2 Instruct 90B (Vision)OpenMeta33.4———
28▼ 2Qwen2 Instruct 72BOpenAlibaba32.6———
29▼ 2Jamba 1.5 LargeOpenAI21 Labs29.3———
30▼ 2Jamba 1.6 LargeOpenAI21 Labs28.8———
31▼ 2Llama 3 Instruct 70BOpenMeta27.5———
32▼ 2Mistral Small (Sep '24)OpenMistral27.4———
33▲ 4Phi-4 Multimodal InstructOpenMicrosoft Azure26.9———
34▼ 2Qwen2.5 Coder Instruct 7BOpenAlibaba26.7———
35▼ 4Mixtral 8x22B InstructOpenMistral26.2———
36▲ 2Phi-4 Mini InstructOpenMicrosoft Azure26.0———
37▼ 4Phi-3 Medium Instruct 14BOpenMicrosoft Azure24.5———
38▼ 3Llama 3.1 Instruct 8BOpenMeta23.7———
39▼ 3Pixtral 12B (2409)OpenMistral23.4———
40▼ 1Ministral 8BOpenMistral22.3———
41NewGemma 2 9BOpenGoogle22.2———
42NewPhi-3 Mini Instruct 3.8BOpenMicrosoft Azure22.1———
43▼ 2Command-R+ (Aug '24)OpenCohere · 3 more variants21.5———
44▼ 2Llama 3 Instruct 8BOpenMeta21.5———
45▼ 2Codestral (May '24)OpenMistral20.4———
46▼ 2Aya Expanse 32BOpenCohere20.1———
47NewLlama 2 Chat 13BOpenMeta19.9———
48▼ 3DBRX InstructOpenDatabricks19.8———
49▼ 3Mistral NeMoOpenMistral19.6———
50▼ 3Llama 3.2 Instruct 3BOpenMeta19.5———

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 “—”.