Skip to content
IFAI

Model rankings

1371 models · 5 boards · 2 sources

Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated July 20, 2026

Best open model
GLM-5.2 (max)Z AI51.1Intelligence
Best value
Qwen3.5 4B (Reasoning)Alibaba335 pts/$Value
Best from China
Kimi K3Kimi57.1Intelligence
Top three by Intelligence
2GPT-5.6 Sol (max)OpenAI58.9
3Kimi K3Kimi57.1
Top three open models
1GLM-5.2 (max)Z AI · #12 overall51.1
2MiniMax-M3MiniMax · #18 overall44.4
3DeepSeek V4 Pro (Reasoning, Max Effort)DeepSeek · #19 overall44.3
This week's changes
kimi-k3-maxCoding · #1New
claude-opus-5-highCoding · #2New
Kimi K3Intelligence · #3New
glm-5.2-maxCoding · #5New
gemini-3.6-flashCoding · #13New
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)ClosedAnthropic59.9$2061113 s
2—GPT-5.6 Sol (max)ClosedOpenAI58.9$1162128 s
3—GPT-5.6 Sol (xhigh)ClosedOpenAI57.7$116042 s
4NewKimi K3ClosedKimi57.1$6.0——
5▼ 1GPT-5.6 Sol (high)ClosedOpenAI55.9$115913 s
6▼ 1Claude Opus 4.8 (Adaptive Reasoning, Max Effort)ClosedAnthropic55.7$105921 s
7▼ 1GPT-5.6 Terra (max)ClosedOpenAI55.0$5.6136140 s
8▼ 1GPT-5.5 (xhigh)ClosedOpenAI54.8$117678 s
9▼ 1Grok 4.5 (high)ClosedSpaceXAI53.8$3.07612 s
10▼ 1GPT-5.6 Sol (medium)ClosedOpenAI53.6$11524.9 s
11▼ 1Claude Opus 4.7 (Adaptive Reasoning, Max Effort)ClosedAnthropic53.5$105020 s
12▼ 1Claude Sonnet 5 (Adaptive Reasoning, Max Effort)ClosedAnthropic53.4$4.083192 s
13▼ 1GPT-5.5 (high)ClosedOpenAI53.1$117125 s
14▼ 1GPT-5.6 Terra (xhigh)ClosedOpenAI51.6$5.612612 s
15▼ 1GPT-5.4 (xhigh)ClosedOpenAI51.4$5.6140112 s
16▼ 1GPT-5.6 Luna (max)ClosedOpenAI51.2$2.3192108 s
17▼ 1Muse Spark 1.1 (xhigh)ClosedMeta50.6$2.01221.3 s
18▼ 1GPT-5.5 (medium)ClosedOpenAI50.4$11724.9 s
19▼ 1Gemini 3.5 Flash (high)ClosedGoogle50.2$3.416921 s
20▼ 1GPT-5.6 Sol (low)ClosedOpenAI49.4$11523.4 s
21▼ 1GPT-5.6 Luna (xhigh)ClosedOpenAI49.1$2.319532 s
22▼ 1GPT-5.6 Terra (high)ClosedOpenAI49.0$5.61202.4 s
23▼ 1Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic47.2$6.054108 s
24▼ 1Gemini 3.1 Pro PreviewClosedGoogle46.5$4.512426 s
25▼ 1GPT-5.6 Luna (high)ClosedOpenAI46.1$2.31896.7 s
26▼ 1Qwen3.7 MaxClosedAlibaba46.0$3.82032.4 s
27▼ 1GPT-5.6 Terra (medium)ClosedOpenAI45.6$5.61191.3 s
28▼ 1Gemini 3.5 Flash (medium)ClosedGoogle45.4$3.416915 s
29▼ 1GPT-5.3 Codex (xhigh)ClosedOpenAI44.3$4.811667 s
30▼ 1Claude Opus 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic43.7$104214 s
31▼ 1GPT-5.5 (low)ClosedOpenAI43.5$11701.6 s
32▼ 1Muse SparkClosedMeta43.1———
33▼ 1Claude Opus 4.7 (Non-reasoning, High Effort)ClosedAnthropic42.7$10461.6 s
34▼ 1GPT-5.2 (xhigh)ClosedOpenAI42.2$4.874114 s
35▼ 1Claude Sonnet 5 (Non-reasoning, High Effort)ClosedAnthropic41.7$4.0651.2 s
36▼ 1GPT-5.6 Sol (Non-reasoning)ClosedOpenAI41.2$11601.0 s
37▼ 1Claude Opus 4.5 (Reasoning)ClosedAnthropic40.8$104713 s
38▼ 1GPT-5.6 Terra (low)ClosedOpenAI40.5$5.61251.2 s
39▼ 1MiMo-V2-ProClosedXiaomi40.3———
40▼ 1GPT-5.2 Codex (xhigh)ClosedOpenAI40.1$4.817452 s
41▼ 1Qwen3.6 Max PreviewClosedAlibaba40.0$2.9383.5 s
42▼ 1GPT-5.4 mini (xhigh)ClosedOpenAI40.0$1.71657.8 s
43▼ 1Grok Build 0.1 0616ClosedSpaceXAI39.8$1.3950.5 s
44▼ 1Qwen3.6 PlusClosedAlibaba39.6$1.1532.6 s
45▼ 1Gemini 3 Pro Preview (high)ClosedGoogle39.6$4.5——
46▼ 1GPT-5.4 (low)ClosedOpenAI39.1$5.6981.4 s
47▼ 1Qwen3.7 PlusClosedAlibaba39.0$0.70532.8 s
48▼ 1JT-4.1 Flash 236B A21BClosedChina Mobile38.8———
49▼ 1GPT-5.4 nano (xhigh)ClosedOpenAI38.2$0.461714.5 s
50▼ 1GLM-5-TurboClosedZ AI38.1———

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 “—”.