Skip to content
IFAI

Model rankings

481 models · 3 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated August 4, 2025

Smartest
Grok 4xAI73.2Intelligence
Best open model
Qwen3 235B A22B 2507 (Reasoning)Alibaba68.9Intelligence
Best value
Gemma 3n E4B InstructGoogle1,120 pts/$Value
Best from China
Qwen3 235B A22B 2507 (Reasoning)Alibaba68.9Intelligence
Top three by Intelligence
1Grok 4xAI73.2
2o3-proOpenAI71.0
3Gemini 2.5 ProGoogle70.5
Top three open models
1Qwen3 235B A22B 2507 (Reasoning)Alibaba · #6 overall68.9
2DeepSeek R1 0528 (May '25)DeepSeek · #8 overall68.3
3GLM-4.5Z AI · #10 overall65.7
This week's changes
GLM-4.5Intelligence · #10New
EXAONE 4.0 32B (Reasoning)Intelligence · #14New
Sonar Reasoning ProIntelligence · #21New
GLM-4.5-AirIntelligence · #24New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—Qwen3 235B A22B 2507 (Reasoning)OpenAlibaba68.9$2.6——
2—DeepSeek R1 0528 (May '25)OpenDeepSeek68.3$0.96——
3NewGLM-4.5OpenZ AI65.7$0.94——
4NewEXAONE 4.0 32B (Reasoning)OpenLG AI Research64.1$0.70——
5NewLlama Nemotron Super 49B v1.5 (Reasoning)OpenNVIDIA63.8$0.000——
6NewQwen3 30B A3B 2507 (Reasoning)OpenAlibaba63.4$0.75——
7▼ 4MiniMax M1 80kOpenMiniMax63.0$0.83——
8▼ 4Qwen3 235B A22B (Reasoning)OpenAlibaba62.3$2.6——
9▼ 4MiniMax M1 40kOpenMiniMax61.4$0.83——
10▼ 4Llama 3.1 Nemotron Ultra 253B v1 (Reasoning)OpenNVIDIA60.8$0.90——
11▼ 4Qwen3 235B A22B 2507 (Non-reasoning)OpenAlibaba60.4$1.2——
12▼ 4DeepSeek R1 (Jan '25)OpenDeepSeek60.2$1.8——
13NewGLM-4.5-AirOpenZ AI59.8$0.43——
14▼ 5Qwen3 32B (Reasoning)OpenAlibaba59.2$2.6——
15▼ 5QwQ 32BOpenAlibaba58.1$0.47——
16▼ 5Kimi K2OpenMoonshot AI57.6$1.1——
17NewQwen3 30B A3B 2507 (Non-reasoning)OpenAlibaba56.6$0.35——
18▼ 6Qwen3 14B (Reasoning)OpenAlibaba55.7$1.3——
19▼ 6Qwen3 30B A3B (Reasoning)OpenAlibaba55.6$0.75——
20▼ 6Magistral SmallOpenMistral54.7$0.75——
21▼ 6Qwen3 Coder 480B A35BOpenAlibaba53.7$3.0——
22▲ 39R1 1776OpenPerplexity53.7$3.5——
23▼ 7DeepSeek V3 0324 (Mar '25)OpenDeepSeek53.2$0.48——
24▼ 7DeepSeek R1 0528 Qwen3 8BOpenDeepSeek52.2$0.065——
25▼ 7DeepSeek R1 Distill Qwen 32BOpenDeepSeek51.9$0.30——
26▼ 7Qwen3 8B (Reasoning)OpenAlibaba51.4$0.66——
27▼ 7Llama 3.3 Nemotron Super 49B v1 (Reasoning)OpenNVIDIA51.2$0.000——
28NewEXAONE 4.0 32BOpenLG AI Research50.8$0.70——
29▼ 8Llama 4 MaverickOpenMeta50.5$0.39——
30▼ 8DeepSeek R1 Distill Qwen 14BOpenDeepSeek49.0$0.20——
31▼ 8DeepSeek R1 Distill Llama 70BOpenDeepSeek48.2$0.72——
32▼ 8Qwen3 4B (Reasoning)OpenAlibaba47.3$0.40——
33▼ 8Reka Flash 3OpenReka AI47.1$0.35——
34▼ 8Qwen3 235B A22BOpenAlibaba47.1$1.2——
35▼ 8DeepSeek V3 (Dec '24)OpenDeepSeek45.6$0.48——
36▼ 8Llama 3.1 Nemotron Nano 4B v1.1 (Reasoning)OpenNVIDIA45.0$0.000——
37▼ 8Qwen3 32BOpenAlibaba43.5$1.2——
38▼ 8Llama 4 ScoutOpenMeta43.0$0.23——
39▼ 8QwQ 32B-PreviewOpenAlibaba42.6$0.67——
40▼ 8Qwen3 30B A3BOpenAlibaba42.6$0.35——
41▼ 8Mistral Small 3.2OpenMistral42.3$0.15——
42▼ 8Llama 3.3 Instruct 70BOpenMeta41.1$0.60——
43▼ 8Qwen3 14BOpenAlibaba41.0$0.61——
44▼ 8Llama 3.1 Instruct 405BOpenMeta40.5$3.3——
45▼ 8Qwen2.5 Instruct 72BOpenAlibaba40.4$0.000——
46▼ 8MiniMax-Text-01OpenMiniMax40.2$0.43——
47▼ 8Phi-4OpenMicrosoft Azure40.2$0.22——
48▼ 8Command AOpenCohere40.0$4.4——
49▼ 8Llama 3.1 Tulu3 405BOpenAllen Institute for AI39.9$0.000——
50▼ 8Llama 3.3 Nemotron Super 49B v1OpenNVIDIA39.3$0.000——

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 “—”.