DW AI LAB

공식 자료 기반의 AI 분석

분석 방법

벤치마크 상세

Artificial Analysis Intelligence Index

최신 종합 지능: 실제 서비스에 어떤 모델을 붙일지 고를 때

공식 리더보드 보기 공식 페이지가 estimated로 표시한 지수는 제외했습니다. 9월 4일 개편된 v4.2 수치이며 이전 v4.1 지수와 직접 비교하지 않습니다.

모든 공개 결과

모델 검색 · 점수순
확인 2026. 09. 07.v4.2 · 2026-09-04Intelligence Index v4.2
55개 결과원문 ↗
순서모델 / 실행 구성점수평가 조건·근거
1Claude Fable 5.1 (Adaptive Reasoning, Max Effort, Default Fallback)Anthropic56.76v4.2 · 2026-09-04
2GPT-6 Astra (max)OpenAI54.66v4.2 · 2026-09-04
3GPT-6 Astra (xhigh)OpenAI54.31v4.2 · 2026-09-04
4Claude Opus 5 (Adaptive Reasoning, Max Effort)Anthropic54.05v4.2 · 2026-09-04
5Claude Opus 5 (Adaptive Reasoning, Xhigh Effort)Anthropic53.37v4.2 · 2026-09-04
6GPT-6 Astra (high)OpenAI53.36v4.2 · 2026-09-04
7Claude Fable 5 (Adaptive Reasoning, Max Effort, Opus 4.8 Fallback)Anthropic53.19v4.2 · 2026-09-04
8Muse Spark 1.3 (max)Meta52.95v4.2 · 2026-09-04
9GPT-6 Astra (medium)OpenAI52.25v4.2 · 2026-09-04
10Claude Opus 5 (Adaptive Reasoning, High Effort)Anthropic52.03v4.2 · 2026-09-04
11Muse Spark 1.3 (xhigh)Meta51.55v4.2 · 2026-09-04
12GPT-5.6 Sol (max)OpenAI51.26v4.2 · 2026-09-04
13Grok 4.6 (high)SpaceXAI50.58v4.2 · 2026-09-04
14Kimi K3 (max)Kimi50.23v4.2 · 2026-09-04
15GPT-5.6 Sol (xhigh)OpenAI49.84v4.2 · 2026-09-04
16Claude Opus 5 (Adaptive Reasoning, Medium Effort)Anthropic49.52v4.2 · 2026-09-04
17GPT-6 Astra (low)OpenAI49.32v4.2 · 2026-09-04
18GLM-5.3 (max)Z AI48.58v4.2 · 2026-09-04
19GPT-5.6 Sol (high)OpenAI48.3v4.2 · 2026-09-04
20GPT-6 Astra (Non-reasoning)OpenAI47.85v4.2 · 2026-09-04
21Gemini 3.8 Flash (high)Google47.07v4.2 · 2026-09-04
22Qwen3.8 MaxAlibaba46.91v4.2 · 2026-09-04
23Muse Spark 1.2 (xhigh)Meta46.84v4.2 · 2026-09-04
24GPT-5.6 Terra (max)OpenAI46.77v4.2 · 2026-09-04
25Qwen3.8 2.4T A95BAlibaba46.74v4.2 · 2026-09-04
26GLM-5.3-FlashZ AI46.22v4.2 · 2026-09-04
27GPT-5.6 Sol (medium)OpenAI45.97v4.2 · 2026-09-04
28Grok 4.5 (high)SpaceXAI45.48v4.2 · 2026-09-04
29Gemini 3.7 Flash (high)Google45.24v4.2 · 2026-09-04
30Claude Sonnet 5 (Adaptive Reasoning, Max Effort)Anthropic45.11v4.2 · 2026-09-04
31GPT-5.6 Terra (xhigh)OpenAI44.39v4.2 · 2026-09-04
32Claude Opus 5 (Adaptive Reasoning, Low Effort)Anthropic43.81v4.2 · 2026-09-04
33GPT-5.6 Luna (max)OpenAI43.44v4.2 · 2026-09-04
34DeepSeek V4 Pro 0813 (Reasoning, Max Effort)DeepSeek42.11v4.2 · 2026-09-04
35GPT-5.6 Luna (xhigh)OpenAI41.58v4.2 · 2026-09-04
36Qwen3.8 27B (xhigh)Alibaba41.41v4.2 · 2026-09-04
37GPT-5.6 Terra (high)OpenAI41.3v4.2 · 2026-09-04
38GPT-5.6 Sol (low)OpenAI40.84v4.2 · 2026-09-04
39DeepSeek V4 Flash 0731 (Reasoning, Max Effort)DeepSeek40.84v4.2 · 2026-09-04
40Gemini 3.6 Flash (high)Google40.29v4.2 · 2026-09-04
41Gemini 3.5 Flash (high)Google39.73v4.2 · 2026-09-04
42K2 Horizon 375B A23BMBZUAI Institute of Foundation Models37.75v4.2 · 2026-09-04
43Gemini 3.1 Pro PreviewGoogle36.65v4.2 · 2026-09-04
44MiniMax-M3MiniMax35.75v4.2 · 2026-09-04
45MiMo-V2.5-ProXiaomi32.59v4.2 · 2026-09-04
46Inkling (xhigh)Thinking Machines32.17v4.2 · 2026-09-04
47Gemini 3.5 Flash-LiteGoogle27.57v4.2 · 2026-09-04
48Ling 3.0 FlashInclusionAI27.36v4.2 · 2026-09-04
49Muse Glimmer (high)Meta24.38v4.2 · 2026-09-04
50Claude 4.5 Haiku (Reasoning)Anthropic22.46v4.2 · 2026-09-04
51Ling 3.0 TinyInclusionAI15.68v4.2 · 2026-09-04
52gpt-oss-120b (high)OpenAI15.57v4.2 · 2026-09-04
53Granite 4.2 8BIBM13.94v4.2 · 2026-09-04
54Mistral Large 3Mistral11.13v4.2 · 2026-09-04
55Granite 4.2 3BIBM10.7v4.2 · 2026-09-04

공개된 숫자만 높은 순으로 표시합니다. 순서가 신뢰구간을 고려한 확정 순위는 아닙니다. 공식 페이지가 estimated로 표시한 지수는 제외했습니다. 9월 4일 개편된 v4.2 수치이며 이전 v4.1 지수와 직접 비교하지 않습니다.

평가 방법 확인 ↗

읽는 법

전체 벤치마크

AA가 독립 측정한 10개 평가를 결합한 v4.2 종합 지수입니다. 속도와 가격은 지수에 합산하지 않습니다.

2026-09-04 개편된 v4.2를 사용합니다. estimated로 표시된 지수는 제외하고 이전 버전과 구분합니다.