Skip to content
IFAI

Model rankings

855 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated December 1, 2025

Smartest
Gemini 3 Pro Preview (high)Google72.9Intelligence
Best open model
Kimi K2 ThinkingMoonshot AI67.0Intelligence
Best value
Gemma 3n E4B InstructGoogle620 pts/$Value
Best from China
Kimi K2 ThinkingMoonshot AI67.0Intelligence
Top three by Intelligence
3GPT-5.1 (high)OpenAI69.7
Top three open models
1Kimi K2 ThinkingMoonshot AI · #6 overall67.0
2MiniMax-M2MiniMax · #13 overall61.4
3gpt-oss-120B (high)OpenAI · #14 overall60.5
This week's changes
Runway Gen-4.5Video · #1New
Claude Opus 4.5 (Reasoning)Intelligence · #2New
FLUX.2 [pro]Image · #2New
Riverflow 2 PreviewImage · #4New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1NewGemini 3 Pro Preview (high)ClosedGoogle72.9$4.5——
2NewClaude Opus 4.5 (Reasoning)ClosedAnthropic69.8$10——
3▼ 1GPT-5.1 (high)ClosedOpenAI69.7$3.4——
4▼ 1GPT-5 Codex (high)ClosedOpenAI68.5$3.4——
5▼ 1GPT-5 (high)ClosedOpenAI68.5$3.4——
6▼ 1Kimi K2 ThinkingOpenMoonshot AI67.0$1.1——
7▼ 1GPT-5 (medium)ClosedOpenAI66.4$3.4——
8▼ 1o3ClosedOpenAI65.5$3.5——
9▼ 1Grok 4ClosedxAI65.3$6.0——
10▼ 1o3-proClosedOpenAI65.3$35——
11▼ 1GPT-5 mini (high)ClosedOpenAI64.3$0.69——
12▼ 1Grok 4.1 Fast (Reasoning)ClosedxAI64.1$0.28——
13▼ 1Claude 4.5 Sonnet (Reasoning)ClosedAnthropic62.7$6.0——
14▼ 1GPT-5 (low)ClosedOpenAI61.8$3.4——
15▼ 1MiniMax-M2OpenMiniMax61.4$0.52——
16▼ 1GPT-5 mini (medium)ClosedOpenAI60.8$0.69——
17▼ 1gpt-oss-120B (high)OpenOpenAI60.5$0.26——
18▼ 1Grok 4 Fast (Reasoning)ClosedxAI60.3$0.28——
19NewClaude Opus 4.5 (Non-reasoning)ClosedAnthropic59.9$10——
20▼ 2Gemini 2.5 ProClosedGoogle59.6$3.4——
21▼ 2o4-mini (high)ClosedOpenAI59.6$1.9——
22▼ 2Claude 4.1 Opus (Reasoning)ClosedAnthropic59.3$30——
23▼ 2DeepSeek V3.1 Terminus (Reasoning)OpenDeepSeek57.7$0.80——
24▼ 2Qwen3 235B A22B 2507 (Reasoning)OpenAlibaba57.5$2.6——
25▼ 2Doubao Seed CodeClosedByteDance Seed57.2$0.41——
26▼ 2Grok 3 mini Reasoning (high)ClosedxAI57.1$0.35——
27▼ 2DeepSeek V3.2 Exp (Reasoning)OpenDeepSeek56.9$0.32——
28▼ 2Claude 4 Sonnet (Reasoning)ClosedAnthropic56.5$6.0——
29▼ 2GLM-4.6 (Reasoning)OpenZ AI56.0$1.0——
30▼ 2Qwen3 Max ThinkingClosedAlibaba55.8$2.4——
31▼ 2Qwen3 MaxClosedAlibaba55.2$2.4——
32▼ 2Claude 4.5 Haiku (Reasoning)ClosedAnthropic54.6$2.0——
33▼ 2Gemini 2.5 Flash Preview (Sep '25) (Reasoning)ClosedGoogle54.4$0.85——
34▼ 2Qwen3 VL 235B A22B (Reasoning)OpenAlibaba54.4$2.6——
35▼ 2Qwen3 Next 80B A3B (Reasoning)OpenAlibaba54.3$1.9——
36▼ 2Claude 4 Opus (Reasoning)ClosedAnthropic54.2$30——
37▼ 2Gemini 2.5 Pro Preview (Mar' 25)ClosedGoogle54.1$3.4——
38▼ 2DeepSeek V3.1 (Reasoning)OpenDeepSeek54.0$0.65——
39▼ 2Gemini 2.5 Pro Preview (May' 25)ClosedGoogle53.2$3.4——
40▼ 2gpt-oss-20B (high)OpenOpenAI52.1$0.10——
41▼ 2Magistral Medium 1.2ClosedMistral52.1$2.8——
42▼ 2DeepSeek R1 0528 (May '25)OpenDeepSeek52.0$2.0——
43▼ 2Qwen3 VL 32B (Reasoning)OpenAlibaba51.9$2.6——
44▼ 2Seed-OSS-36B-InstructOpenByteDance Seed51.6$0.30——
45▼ 2Apriel-v1.5-15B-ThinkerOpenServiceNow51.6$0.000——
46▼ 2GLM-4.5 (Reasoning)OpenZ AI51.3$0.98——
47▼ 2Gemini 2.5 Flash (Reasoning)ClosedGoogle51.2$0.85——
48▼ 2GPT-5 nano (high)ClosedOpenAI51.0$0.14——
49▼ 2o3-mini (high)ClosedOpenAI50.8$1.9——
50▼ 2Kimi K2 0905OpenMoonshot AI50.4$1.4——

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 “—”.