Skip to content
IFAI

Model rankings

1092 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated March 25, 2026

Smartest
Gemini 3.1 Pro PreviewGoogle57.2Intelligence
Best open model
GLM-5 (Reasoning)Z AI49.8Intelligence
Best value
Qwen3.5 9B (Reasoning)Alibaba288 pts/$Value
Best from China
GLM-5 (Reasoning)Z AI49.8Intelligence
Top three by Intelligence
2GPT-5.4 (xhigh)OpenAI57.2
Top three open models
1GLM-5 (Reasoning)Z AI · #7 overall49.8
2Kimi K2.5 (Reasoning)Kimi · #16 overall46.8
3Qwen3.5 397B A17B (Reasoning)Alibaba · #19 overall45.1
This week's changes
MiniMax-M2.7Intelligence · #9New
MiMo-V2-ProIntelligence · #10New
GPT-5.4 mini (xhigh)Intelligence · #14New
GLM-5-TurboIntelligence · #17New
GPT-5.4 nano (xhigh)Intelligence · #22New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—Gemini 3.1 Pro PreviewClosedGoogle57.2$4.5——
2—GPT-5.4 (xhigh)ClosedOpenAI57.2$5.6——
3—GPT-5.3 Codex (xhigh)ClosedOpenAI54.0$4.8——
4—Claude Opus 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic53.0$10——
5—Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic51.7$6.0——
6—GPT-5.2 (xhigh)ClosedOpenAI51.3$4.8——
7—GLM-5 (Reasoning)OpenZ AI49.8$1.6——
8—Claude Opus 4.5 (Reasoning)ClosedAnthropic49.7$10——
9NewMiniMax-M2.7ClosedMiniMax49.6$0.52——
10NewMiMo-V2-ProClosedXiaomi49.2$1.5——
11▼ 2GPT-5.2 Codex (xhigh)ClosedOpenAI49.0$4.8——
12▼ 2Grok 4.20 Beta 0309 (Reasoning)ClosedxAI48.5$3.0——
13▼ 2Gemini 3 Pro Preview (high)ClosedGoogle48.4$4.5——
14NewGPT-5.4 mini (xhigh)ClosedOpenAI48.1$1.7——
15▼ 3GPT-5.1 (high)ClosedOpenAI47.7$3.4——
16▼ 3Kimi K2.5 (Reasoning)OpenKimi46.8$1.2——
17NewGLM-5-TurboClosedZ AI46.8$0.000——
18▼ 4GPT-5.2 (medium)ClosedOpenAI46.6$4.8——
19▼ 4Claude Opus 4.6 (Non-reasoning, High Effort)ClosedAnthropic46.5$10——
20▼ 4Gemini 3 Flash Preview (Reasoning)ClosedGoogle46.4$1.1——
21▼ 4Qwen3.5 397B A17B (Reasoning)OpenAlibaba45.1$1.4——
22▼ 4GPT-5 (high)ClosedOpenAI44.6$3.4——
22▼ 4GPT-5 Codex (high)ClosedOpenAI44.6$3.4——
24NewGPT-5.4 nano (xhigh)ClosedOpenAI44.4$0.46——
25▼ 5Claude Sonnet 4.6 (Non-reasoning, High Effort)ClosedAnthropic44.4$6.0——
26Newmimo-v2-omniClosedXiaomi43.4$0.000——
27▼ 6GPT-5.1 Codex (high)ClosedOpenAI43.1$3.4——
28▼ 6Claude Opus 4.5 (Non-reasoning)ClosedAnthropic43.1$10——
29▼ 6Claude 4.5 Sonnet (Reasoning)ClosedAnthropic43.0$6.0——
30▼ 6Claude Sonnet 4.6 (Non-reasoning, Low Effort)ClosedAnthropic42.6$6.0——
31▼ 6GLM-4.7 (Reasoning)OpenZ AI42.1$1.0——
32▼ 6Qwen3.5 27B (Reasoning)OpenAlibaba42.1$0.83——
33▼ 6GPT-5 (medium)ClosedOpenAI42.0$3.4——
34▲ 46Claude 4.1 Opus (Reasoning)ClosedAnthropic42.0$30——
35▼ 7MiniMax-M2.5OpenMiniMax41.9$0.52——
36▼ 7DeepSeek V3.2 (Reasoning)OpenDeepSeek41.7$0.32——
37▼ 7Qwen3.5 122B A10B (Reasoning)OpenAlibaba41.6$1.1——
38▼ 7Grok 4ClosedxAI41.5$6.0——
39▼ 7MiMo-V2-Flash (Feb 2026)OpenXiaomi41.5$0.15——
40▼ 7Gemini 3 Pro Preview (low)ClosedGoogle41.3$4.5——
41▼ 7GPT-5 mini (high)ClosedOpenAI41.2$0.69——
42▼ 7Kimi K2 ThinkingOpenKimi40.9$1.1——
43▼ 7o3-proClosedOpenAI40.7$35——
44▼ 7GLM-5 (Non-reasoning)OpenZ AI40.6$1.6——
45▼ 7Qwen3.5 397B A17B (Non-reasoning)OpenAlibaba40.1$1.4——
46▼ 7Qwen3 Max ThinkingClosedAlibaba39.9$2.4——
47▼ 7MiniMax-M2.1OpenMiniMax39.4$0.52——
48▼ 7MiMo-V2-Flash (Reasoning)OpenXiaomi39.2$0.15——
49▼ 7GPT-5 (low)ClosedOpenAI39.2$3.4——
50▲ 59Claude 4 Opus (Reasoning)ClosedAnthropic39.0$30——

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 “—”.