Skip to content
IFAI

Model rankings

1470 models · 5 boards · 2 sources

‹
Week of August 31, 2026
October 12, 2026127 changesOctober 5, 202634 changesSeptember 28, 202617 changesSeptember 21, 2026115 changesSeptember 14, 2026174 changesSeptember 7, 2026144 changesAugust 31, 202620 changesAugust 24, 202615 changesAugust 17, 202633 changesAugust 10, 2026116 changesAugust 3, 202613 changesJuly 27, 202639 changes
›
Artificial AnalysisIntelligence IndexArenaText arena

Intelligence Index · A composite of several evaluations · Updated August 31, 2026

Smartest
Claude Opus 5 (Adaptive Reasoning, Max Effort)Anthropic63.1Intelligence
Best open model
Kimi K3 (max)Kimi59.7Intelligence
Best value
Hy3-preview (Reasoning)Tencent345 pts/$Value
Best from China
Kimi K3 (max)Kimi59.7Intelligence
Top three by Intelligence
1Claude Opus 5 (Adaptive Reasoning, Max Effort)Anthropic63.1
2Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic62.1
3GPT-5.6 Sol (max)OpenAI60.9
Top three open models
1Kimi K3 (max)Kimi · #5 overall59.7
2GLM-5.3 (max)Z AI · #6 overall59.5
3Qwen3.8 2.4T A95BAlibaba · #8 overall57.7
This week's changes
gpt-6-astra-maxCoding · #1New
Wan 3.0Video · #1New
claude-fable-5.1-maxCoding · #2New
MAI-Image-2.6-PreviewImage · #2New
GLM-5.3-FlashIntelligence · #9New
AllOpenClosed
All regionsUSChina
One row per modelAll variants
SortIntelligenceValueSpeedLowest price
RankChangeModelIntelligencePrice $/MSpeed tok/sFirst token
1—Claude Opus 5 (Adaptive Reasoning, Max Effort)ClosedAnthropic · 4 more variants63.1$105362 s
2—Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)ClosedAnthropic62.1$2067100 s
3—GPT-5.6 Sol (max)ClosedOpenAI · 5 more variants60.9$8.079116 s
4—Claude Opus 4.8 (Adaptive Reasoning, Max Effort)ClosedAnthropic57.3$105742 s
5—Muse Spark 1.2 (xhigh)ClosedMeta56.8$2.0——
6—GPT-5.6 Terra (max)ClosedOpenAI · 5 more variants56.6$4.5122137 s
7—GPT-5.5 (xhigh)ClosedOpenAI · 4 more variants56.3$119445 s
8—Gemini 3.7 Flash (high)ClosedGoogle · 2 more variants56.0$1.53159.2 s
9—Claude Sonnet 5 (Adaptive Reasoning, Max Effort)ClosedAnthropic · 1 more variant55.3$4.075200 s
10—Claude Opus 4.7 (Adaptive Reasoning, Max Effort)ClosedAnthropic · 1 more variant55.0$104718 s
11—Muse Spark 1.1 (xhigh)ClosedMeta53.2$2.01971.4 s
12—GPT-5.4 (xhigh)ClosedOpenAI · 2 more variants53.1$5.616498 s
13—GPT-5.6 Luna (max)ClosedOpenAI · 5 more variants52.3$0.45130147 s
14—Gemini 3.5 Flash (high)ClosedGoogle · 2 more variants52.0$3.421316 s
15—Gemini 3.6 Flash (high)ClosedGoogle51.6$1.517317 s
16—Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic · 2 more variants48.4$6.068121 s
17—Gemini 3.1 Pro PreviewClosedGoogle47.7$4.511630 s
18—GPT-5.3 Codex (xhigh)ClosedOpenAI45.5$4.813241 s
19—Claude Opus 4.6 (Adaptive Reasoning, Max Effort)ClosedAnthropic · 1 more variant44.9$104021 s
20—Muse SparkClosedMeta44.3———
21—GPT-5.2 (xhigh)ClosedOpenAI · 2 more variants43.3$4.883150 s
22—Inkling (xhigh)OpenThinking Machines42.3$1.8772.4 s
23—Claude Opus 4.5 (Reasoning)ClosedAnthropic · 1 more variant41.9$104715 s
24—GPT-5.2 Codex (xhigh)ClosedOpenAI41.2$4.8——
25—Inkling SmallOpenThinking Machines41.2$0.52482.6 s
26—GPT-5.4 mini (xhigh)ClosedOpenAI · 2 more variants40.9$1.71895.1 s
27—Gemini 3 Pro Preview (high)ClosedGoogle · 1 more variant40.6$4.5——
28—GPT-5.4 nano (xhigh)ClosedOpenAI · 2 more variants39.7$0.461835.2 s
29—Gemini 3 Flash Preview (Reasoning)ClosedGoogle · 1 more variant38.7$1.11796.0 s
30—Nemotron 3 Ultra 550B A55B (Reasoning)OpenNVIDIA38.3$1.1705.3 s
31—GPT-5.1 (high)ClosedOpenAI · 1 more variant37.5$3.410547 s
32—Gemini 3.5 Flash-LiteClosedGoogle37.4$0.853589.2 s
33—Claude 4.5 Sonnet (Reasoning)ClosedAnthropic · 1 more variant37.4$6.04313 s
34—GPT-5 Codex (high)ClosedOpenAI37.0$3.4——
35—GPT-5.1 Codex (high)ClosedOpenAI35.6$3.4——
36—GPT-5 (high)ClosedOpenAI · 4 more variants35.3$3.410375 s
37—Muse Glimmer (high)OpenMeta35.1$0.581020.9 s
38—Claude 4.1 Opus (Reasoning)ClosedAnthropic · 1 more variant34.5$30——
39—GPT-5.5 Instant (May 2026)ClosedOpenAI · 1 more variant34.3$11——
40—o3-proClosedOpenAI33.3$35——
41—Claude 4 Opus (Reasoning)ClosedAnthropic · 1 more variant31.7$30——
42—GPT-5 mini (medium)ClosedOpenAI · 2 more variants31.6$0.6911512 s
43—GPT-5.1 Codex mini (high)ClosedOpenAI31.3$0.69——
44—o3ClosedOpenAI31.1$3.51217.1 s
45—Claude 4.5 Haiku (Reasoning)ClosedAnthropic · 1 more variant29.9$2.09221 s
46—Claude 4 Sonnet (Reasoning)ClosedAnthropic · 1 more variant29.8$6.0——
47—Gemma 4 31B (Reasoning)OpenGoogle · 1 more variant29.7$0.000361.1 s
48—Claude 3.7 Sonnet (Reasoning)ClosedAnthropic · 1 more variant27.6———
49—Gemma 4 26B A4B (Reasoning)OpenGoogle · 1 more variant26.1$0.18——
50—o4-mini (high)ClosedOpenAI26.1$1.916323 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 “—”.