Artificial Analysis Intelligence Index
A composite AI model intelligence scoring index published by the Artificial Analysis platform. It aggregates weighted results from multiple authoritative benchmarks, including MMLU, HumanEval, MATH, and GPQA, to evaluate models across reasoning, coding, mathematics, and knowledge question-answering capabilities.
Core Facts
There is a gap between benchmark scores and real user experience, and some vendors engage in targeted optimization for specific benchmarks (score gaming) as well as training data contamination that inflates scores.
Artificial Analysis gave Argon an intelligence index score of 53, on par with GPT-6 Astra.
The Artificial Analysis Intelligence Index provides a unified intelligence score by integrating multiple evaluations including reasoning, coding, mathematics, and knowledge Q&A.
Timeline (last 90 days)
The Artificial Analysis Intelligence Index's evaluation methodology, prompt templates and scoring scripts are publicly available and reproducible, and its update cadence is relatively stable.
Gemini 4 Carbon scored 59 on the Artificial Analysis Intelligence Index.
The aggregated index of the Artificial Analysis Intelligence Index has limitations: the choice of weights can amplify or mask differences in specific capabilities, and benchmark questions may already be covered by models' training data, so a higher ranking does not equate to better performance in real business scenarios.
On the Artificial Analysis Intelligence Index (v5.0), Claude Opus 5.5 and GPT-6 Astra rank immediately behind Gemini 4 Carbon.
According to Artificial Analysis's evaluation, Gemini 4 Carbon ranks first on its intelligence index, just one point ahead of Cloud Opus 5.5.
The Intelligence Index is a composite score obtained by aggregating results from multiple evaluation tasks, normalizing them and weighting them.
When making cross-version comparisons, users should first check the official methodology documentation and version changelogs rather than simply comparing historical numbers.
Version upgrades of evaluation benchmarks are often accompanied by adjustments to scoring conventions such as normalization methods, task weights or aggregation algorithms, so the same model's scores under the old and new versions cannot be directly subtracted.
Kimi K3.1 scored 55 on the Artificial Analysis composite benchmark.
According to Artificial Analysis's Intelligence Index, Opus scores 58, while Astra and Argon both score 53.
15 more timeline events
All Facts (20)
There is a gap between benchmark scores and real user experience, and some vendors engage in targeted optimization for specific benchmarks (score gaming) as well as training data contamination that inflates scores.
90%VerifiedArtificial Analysis gave Argon an intelligence index score of 53, on par with GPT-6 Astra.
80%VerifiedThe Artificial Analysis Intelligence Index provides a unified intelligence score by integrating multiple evaluations including reasoning, coding, mathematics, and knowledge Q&A.
75%VerifiedArtificial Analysis是一个独立的AI模型基准测试平台，综合编程能力、数学推理、语言理解、知识问答等多个维度评估模型表现
70%VerifiedArtificial Analysis Intelligence Index整合了MMLU、HumanEval、MATH、GPQA等多个基准测试的加权结果
65%VerifiedArtificial Analysis released version v4.2 of the Intelligence Index.
65%Partially VerifiedAccording to Artificial Analysis's evaluation, Gemini 4 Carbon ranks first on its intelligence index, just one point ahead of Cloud Opus 5.5.
80%Partially VerifiedGemini 4 Carbon scored 59 on the Artificial Analysis Intelligence Index.
75%UnverifiedThe Artificial Analysis Index (AA) is published by the independent evaluation organization Artificial Analysis, aggregating benchmarks such as MMLU, GPQA, HumanEval, and MATH to calculate a composite score from 0 to 100.
70%Partially VerifiedKimi K3.1 scored 55 on the Artificial Analysis composite benchmark.
65%UnverifiedThe Artificial Analysis Intelligence Index's evaluation methodology, prompt templates and scoring scripts are publicly available and reproducible, and its update cadence is relatively stable.
50%UnverifiedThe aggregated index of the Artificial Analysis Intelligence Index has limitations: the choice of weights can amplify or mask differences in specific capabilities, and benchmark questions may already be covered by models' training data, so a higher ranking does not equate to better performance in real business scenarios.
50%UnverifiedOn the Artificial Analysis Intelligence Index (v5.0), Claude Opus 5.5 and GPT-6 Astra rank immediately behind Gemini 4 Carbon.
50%UnverifiedWhen making cross-version comparisons, users should first check the official methodology documentation and version changelogs rather than simply comparing historical numbers.
50%UnverifiedThe Intelligence Index is a composite score obtained by aggregating results from multiple evaluation tasks, normalizing them and weighting them.
50%UnverifiedVersion upgrades of evaluation benchmarks are often accompanied by adjustments to scoring conventions such as normalization methods, task weights or aggregation algorithms, so the same model's scores under the old and new versions cannot be directly subtracted.
50%UnverifiedAccording to Artificial Analysis's Intelligence Index, Opus scores 58, while Astra and Argon both score 53.
50%UnverifiedThe Artificial Analysis Intelligence Index is a public leaderboard maintained by the independent evaluation firm Artificial Analysis, covering about 200-plus models.
50%UnverifiedAlibaba claims the related system's score on the Artificial Analysis Intelligence Index rose from 40 to 45 within 30 days.
50%UnverifiedOn the Artificial Analysis Intelligence Index, Haiku 4.5 scored 17 points, while Haiku 5.5 jumped to 43 points, second only to Kimi K3's 44 points.
50%