Skip to content
IFAI

Model rankings

1278 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated June 19, 2026

Best open model
GLM-5.2 (max)Z AI51.1Intelligence
Best value
HyperNova 60B 2605Multiverse Computing341 pts/$Value
Best from China
GLM-5.2 (max)Z AI51.1Intelligence
Top three open models
1GLM-5.2 (max)Z AI · #6 overall51.1
2MiniMax-M3MiniMax · #11 overall44.4
3DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek · #12 overall44.3
This week's changes
glm-5.2 (max)Coding · #2New
GLM-5.2 (max)Intelligence · #6New
seed-2.1-pro-previewCoding · #6New
kimi-k2.7-codeCoding · #17New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)ClosedAnthropic59.9$20——
2—Claude Opus 4.8 (Adaptive Reasoning, Max Effort)ClosedAnthropic55.7$105921 s
3—GPT-5.5 (xhigh)ClosedOpenAI54.8$115684 s
4▲ 1Claude Opus 4.7 (Adaptive Reasoning, Max Effort)ClosedAnthropic53.5$105318 s
5▼ 1GPT-5.5 (high)ClosedOpenAI53.1$115640 s
6▲ 1GPT-5.4 (xhigh)ClosedOpenAI51.4$5.6146115 s
7NewGLM-5.2 (max)OpenZ AI51.1$2.2992.1 s
8▲ 2Gemini 3.5 Flash (high)ClosedGoogle50.2$3.416118 s
9▲ 13Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic47.2$6.057106 s
10▼ 2GPT-5.5 (medium)ClosedOpenAI47.1$115210 s
11▼ 5Gemini 3.1 Pro PreviewClosedGoogle46.5$4.512319 s
12▼ 3Qwen3.7 MaxClosedAlibaba46.0$3.81092.7 s
13▼ 2Gemini 3.5 Flash (medium)ClosedGoogle45.4$3.416217 s
14▼ 2MiniMax-M3OpenMiniMax44.4$0.52602.7 s
15▲ 8DeepSeek V4 Pro (Reasoning, Max Effort)OpenDeepSeek44.3$0.54751.8 s
16▼ 1GPT-5.3 Codex (xhigh)ClosedOpenAI44.3$4.87287 s
17▲ 1Claude Opus 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic43.7$104420 s
18▲ 1Muse SparkClosedMeta43.1———
19▼ 6Kimi K2.6OpenKimi42.8$1.7442.5 s
20—Claude Opus 4.7 (Non-reasoning, High Effort)ClosedAnthropic42.7$10461.2 s
21▼ 7MiMo-V2.5-ProOpenXiaomi42.2$0.54502.6 s
22▲ 3GPT-5.2 (xhigh)ClosedOpenAI42.2$4.870133 s
23NewKimi K2.7 CodeOpenKimi41.9$1.7562.3 s
24▲ 2GPT-5.5 (low)ClosedOpenAI41.7$11531.7 s
25▲ 3DeepSeek V4 Pro (Reasoning, High Effort)OpenDeepSeek40.8$0.54691.8 s
26▲ 4Claude Opus 4.5 (Reasoning)ClosedAnthropic40.8$105514 s
27▲ 6MiMo-V2-ProClosedXiaomi40.3$1.5512.6 s
28▲ 18DeepSeek V4 Flash (Reasoning, Max Effort)OpenDeepSeek40.3$0.181071.3 s
29▼ 5GLM-5.1 (Reasoning)OpenZ AI40.2$2.2791.5 s
30▲ 4GPT-5.2 Codex (xhigh)ClosedOpenAI40.1$4.813838 s
31▲ 4MiMo-V2.5OpenXiaomi40.1$0.18782.5 s
32▼ 11Qwen3.6 Max PreviewClosedAlibaba40.0$2.9423.4 s
33▲ 3GPT-5.4 mini (xhigh)ClosedOpenAI40.0$1.71758.6 s
34▼ 7Qwen3.6 PlusClosedAlibaba39.6$1.1502.7 s
35▲ 4Gemini 3 Pro Preview (high)ClosedGoogle39.6$4.5——
36▼ 7GLM-5 (Reasoning)OpenZ AI39.5$1.6751.7 s
37▲ 3GPT-5.4 (low)ClosedOpenAI39.1$5.9951.6 s
38▼ 22Qwen3.7 PlusClosedAlibaba39.0$0.59512.6 s
39▲ 2GPT-5.1 (high)ClosedOpenAI38.9$3.49343 s
40▲ 16GPT-5.4 nano (xhigh)ClosedOpenAI38.2$0.461575.0 s
41▼ 10MiniMax-M2.7OpenMiniMax38.1$0.52442.1 s
42▲ 1Kimi K2.5 (Reasoning)OpenKimi38.1$1.2553.0 s
43▲ 1GLM-5-TurboClosedZ AI38.1———
44▲ 1GPT-5.2 (medium)ClosedOpenAI38.0$4.8——
45▲ 2Claude Opus 4.6 (Non-reasoning, High Effort)ClosedAnthropic37.8$10402.3 s
46▼ 4Nemotron 3 Ultra 550B A55B (Reasoning)OpenNVIDIA37.8$1.21701.1 s
47▲ 1Gemini 3 Flash Preview (Reasoning)ClosedGoogle37.8$1.11666.9 s
48▼ 31Grok 4.3 (high)ClosedxAI37.6$1.615211 s
49—DeepSeek V4 Flash (Reasoning, High Effort)OpenDeepSeek37.4$0.18——
50—Qwen3.6 27B (Reasoning)OpenAlibaba37.1$1.4553.8 s

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 “—”.