独立 AI 分析

了解 AI 格局，为你的使用场景选择最佳模型和服务商

亮点

智能

Artificial Analysis Intelligence Index · Higher is better

速度

Output tokens per second · Higher is better

每任务成本

Weighted average cost (USD) per Intelligence Index task · Lower is better

Optima

Artificial Analysis 出品

新

构建你自己的定制基准测试

模型推荐器

根据你对智能、速度和成本的优先级，获取经过优化的个性化模型推荐

更新日志
新增语言模型评测 · 8月26日
Agnes 2.5 Pro BetaAgnes 2.5 Pro Beta
新增语言模型评测 · 8月26日
Granite 4.2 8BGranite 4.2 8B
新增语言模型评测 · 8月26日
Granite 4.2 3BGranite 4.2 3B
新增语言模型评测 · 8月26日
GLM-5.3-FlashGLM-5.3-Flash
新文章发布 · 8月24日
Announcing the Speech Agent Arena: Compare Speech agents in real world conversations
新增语言模型评测 · 8月24日
DeepSeek V4 Flash Vision (Reasoning, Max Effort)DeepSeek V4 Flash Vision (Reasoning, Max Effort)
新增语言模型评测 · 8月24日
Qwen3.8 27B (Non-reasoning)Qwen3.8 27B (Non-reasoning)
新增语言模型评测 · 8月21日
Grok 4.6 (low)Grok 4.6 (low)
新增语言模型评测 · 8月21日
Grok 4.6 (medium)Grok 4.6 (medium)
新增语言模型评测 · 8月21日
Grok 4.6 (xhigh)Grok 4.6 (xhigh)
新增语言模型评测 · 8月21日
Qwen3.8 27B (low)Qwen3.8 27B (low)
新增语言模型评测 · 8月21日
Qwen3.8 27B (medium)Qwen3.8 27B (medium)
新增语言模型评测 · 8月21日
LFM2.5-2.6BLFM2.5-2.6B
方法论更新 · 8月20日
Artificial AnalysisCoding Agent Index methodology update (v1.4)
新功能发布 · 8月19日
Artificial AnalysisAbout Us
新增语言模型评测 · 8月19日
G9v3-39A5BG9v3-39A5B
新文章发布 · 8月18日
Announcing the Artificial Analysis Search Index: Same Agent, Different Search
新增语言模型评测 · 8月18日
GLM-5.3 (max)GLM-5.3 (max)
新增语言模型评测 · 8月17日
Qwen3.8 27B (xhigh)Qwen3.8 27B (xhigh)
新文章发布 · 8月13日
Gemini 3.7 Flash: On the Intelligence vs. Time per Task Pareto frontier查看更多

智能

根据我们的独立评测衡量领先 AI 模型的智能

Artificial Analysis Intelligence Index

Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Artificial Analysis Intelligence Index by Open Weights / Proprietary

Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

Indicates whether the model weights are available. Models are labelled as 'Commercial Use Restricted' if commercial use is limited by conditions, and as 'Non-commercial' if the license prohibits commercial use.

Cost per Intelligence Index Task

Weighted average cost (USD) per Artificial Analysis Intelligence Index task, segmented by token type. Lower is better
Reasoning models are indicated by a lightbulb icon

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Intelligence Index vs. Cost per Intelligence Index Task

Artificial Analysis Intelligence Index · Weighted average cost (USD) per Artificial Analysis Intelligence Index task
Most attractive quadrant
Pareto line
Reasoning models are indicated by a lightbulb icon

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

前沿语言模型的智能变化

Artificial Analysis Intelligence Index v4.1.1 incorporates 9 evaluations: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

编程智能体指数

领先编程智能体在端到端软件工程任务中的表现、成本和执行时间

Artificial Analysis Coding Agent Index

Composite average pass@1 across DeepSWE, Terminal-Bench v2.1, and SWE-Atlas-QnA · Higher is better

图像与视频

图像竞技场和视频竞技场排行榜中的顶尖模型，包含 95% 置信区间

文生图排行榜

来自图像竞技场盲测偏好投票的 Elo 分数。在此查看完整排行榜。

语音

文本转语音竞技场、语音转文本和语音到语音评测中的顶尖模型

Text to Speech Arena Leaderboard

Elo scores from blind preference votes in our Text to Speech Arena · See the full leaderboard here.

Relative Elo score of the models as determined by responses from users in Artificial Analysis' Speech Arena. Some models may not be shown due to not yet having enough votes.

能力指数

查看更多

衡量模型在特定能力和行业中的表现

Artificial Analysis Agentic Index

Measures performance in agentic workflows, focusing on behaviors like tool use, planning, autonomy, and complex problem solving.
Reasoning models are indicated by a lightbulb icon

基准测试

Intelligence Evaluations

Intelligence evaluations measured independently by Artificial Analysis · Higher is better
See more
GDPval-AA v2

Agentic real-world work tasks, (Elo-500)/2000

𝜏³-Banking

Agentic tool use

Terminal-Bench v2.1

Agentic coding & terminal use

SciCode

Coding

Humanity's Last Exam

Reasoning & knowledge

GPQA Diamond

Scientific reasoning

CritPt

Physics reasoning

AA-Omniscience Accuracy

Knowledge

AA-Omniscience Non-Hallucination Rate

1 - hallucination rate

AA-LCR

Long context reasoning

AA-Briefcase

Agentic knowledge work, Elo

AutomationBench-AA

Agentic SaaS workflows

Harvey LAB-AA

Legal agentic work, criterion pass rate

EnterpriseOps-Gym-AA

Agentic business operations

AA-AnalystAgentNew

Quantitative analysis on spreadsheets & documents

IFBench

Instruction following

APEX-Agents-AA

Long-horizon agentic tasks

ITBench-AA

Kubernetes incident root-cause analysis

MMMU-Pro

Visual reasoning

Reasoning models are indicated by a lightbulb icon

While model intelligence generally translates across use cases, specific evaluations may be more relevant for certain use cases.

Artificial Analysis Intelligence Index v4.1.1 includes: GDPval-AA v2, 𝜏³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, AA-LCR. See Intelligence Index methodology for further details, including a breakdown of each evaluation and how we run them.

AA-Briefcase

AA-Briefcase 是一项面向长周期知识工作的前沿智能体评测，通过要求交付电子表格、演示文稿和备忘录等成果的真实商业工作流来测试智能体

AA-Briefcase Elo

AA-Briefcase is an agentic knowledge work benchmark developed by Artificial Analysis. AA-Briefcase Elo is a combined metric that aggregates rubric pass rate, analytical quality Elo and presentation Elo · Higher is better
Reasoning models are indicated by a lightbulb icon

AA-Briefcase Elo is a combined metric that aggregates analytical quality Elo, presentation Elo, and rubric pass rate, with rubric performance converted into Elo via synthetic head-to-head matches. Elo and 95% confidence interval bounds are clamped at 0.

AA-AnalystAgent

AA-AnalystAgent 是一项针对真实电子表格和文档进行端到端定量分析的基准测试，考察业务分析师和数据分析师日常所做的工作

AA-AnalystAgent pass^5

Share of end-to-end quantitative analysis tasks solved on all five attempts · Higher is better
Reasoning models are indicated by a lightbulb icon

Share of AA-AnalystAgent questions answered correctly on all five attempts. AA-AnalystAgent is Artificial Analysis' benchmark for end-to-end quantitative analysis on real-world spreadsheets and documents; every question is run five independent times, so pass^5 measures how reliably a model reproduces a correct answer rather than how often it reaches one.

AA-Omniscience

AA-Omniscience 是一项知识与幻觉基准测试，奖励准确回答、惩罚不当猜测，并全面展示不同模型在各领域生成事实可靠内容的能力

AA-Omniscience Index

AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct.
Reasoning models are indicated by a lightbulb icon

AA-Omniscience Index (higher is better) measures knowledge reliability and hallucination. It rewards correct answers, penalizes hallucinations, and has no penalty for refusing to answer. Scores range from -100 to 100, where 0 means as many correct as incorrect answers, and negative scores mean more incorrect than correct.

GDPval-AA v2

GDPval-AA v2 在广泛职业中使用具有真实经济价值的任务评测 AI 模型

GDPval-AA v2 Leaderboard

Elo rating for performance on real-world work tasks · Anchored to a human baseline of 1,000 · Higher is better
Human Baseline (1,000)
Reasoning models are indicated by a lightbulb icon

开放性指数

Artificial Analysis 开放性指数根据模型各个组成部分的可获取性和透明度，评估模型的“开放”程度。

Artificial Analysis Openness Index: Components

Openness Index underlying score contribution by components, up to a maximum of 18 (higher is more open)
Reasoning models are indicated by a lightbulb icon

Artificial Analysis Openness Index vs. Artificial Analysis Intelligence Index

Most attractive quadrant
Pareto line

输出 Token

根据我们的独立评测统计领先 AI 模型的输出 token 数

Output Tokens per Intelligence Index Task

Weighted average number of output tokens used to run one task in the Artificial Analysis Intelligence Index
Reasoning models are indicated by a lightbulb icon

The number of tokens required per Intelligence Index task. This is calculated by multiplying the output tokens per eval by the relative weights of each benchmark in the Intelligence Index, then dividing by task count (excluding repeats).

成本

根据我们的独立评测分析领先 AI 模型的价格和实际成本

Cost per Intelligence Index Task

Weighted average cost (USD) per Artificial Analysis Intelligence Index task, segmented by token type. Lower is better
Reasoning models are indicated by a lightbulb icon

Weighted average cost per Intelligence Index task. Each evaluation’s cost is calculated from input, cache hit, cache write, reasoning, and answer token prices, divided by task count, and weighted by its Intelligence Index weight.

Cost to Run Artificial Analysis Intelligence Index

Cost (USD) to run all evaluations in the Artificial Analysis Intelligence Index
Reasoning models are indicated by a lightbulb icon

The cost to run the evaluations in the Artificial Analysis Intelligence Index, calculated using the model's input, cache hit, cache write, reasoning, and answer token prices and the number of tokens used across evaluations (excluding repeats).

Pricing: Cache Hit, Input, and Output

Price (USD per M Tokens)
Reasoning models are indicated by a lightbulb icon

Price per token for cached prompts (previously processed), typically offering a significant discount compared to regular input price, represented as USD per million tokens. The values shown here are the cache hit price; cache write and cache storage are billed separately and vary by provider — see "Cache pricing by provider" for detail.

速度与延迟

比较第一方 API 的性能

Output Speed

Output tokens per second · Higher is better
Reasoning models are indicated by a lightbulb icon

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).

Time per Intelligence Index Task

Weighted average decode time (minutes) per task; excludes TTFT and overhead time · Lower is better
Reasoning models are indicated by a lightbulb icon

The weighted average time (seconds) per Artificial Analysis Intelligence Index task. This is calculated by dividing output tokens per task by output speed, weighted by the relative weights of each benchmark in the Intelligence Index.

服务商

Endpoint Accuracy Index: gpt-oss-120b (high)

v1.0 · Composite of BFCL v4-500, HLE-250 and AA-LCR-25 run against each provider endpoint · Percentage of the reference endpoint, with 95% confidence interval · Higher is better

Composite measure of how much of a model's accuracy a given provider endpoint preserves, from re-running BFCL v4-500, HLE-250 and AA-LCR-25 against that endpoint. Where a self-hosted reference endpoint exists, scores are expressed as a percentage of that reference (100 = matches reference); lower scores indicate accuracy lost to quantisation, sampling defaults, or other endpoint-side configuration. Scores are point-in-time snapshots. Methodology.

Output Speed vs. Price: gpt-oss-120b (high)

Output tokens per second · USD per 1M tokens (blended) · 10,000 input tokens
Most attractive quadrant
Pareto line
Reasoning models are indicated by a lightbulb icon

Smaller, emerging providers are offering high output speed and at competitive prices.

价格（缓存命中、输入与输出）：gpt-oss-120b (high)

Price (USD per M Tokens) · Lower is better · 10,000 input tokens

Price per token for cached prompts (previously processed), typically offering a significant discount compared to regular input price, represented as USD per million tokens. The values shown here are the cache hit price; cache write and cache storage are billed separately and vary by provider — see "Cache pricing by provider" for detail.

Output Speed: gpt-oss-120b (high)

Output speed: output tokens per second · 10,000 input tokens

Tokens per second received while the model is generating tokens (ie. after first chunk has been received from the API for models which support streaming).

Figures represent performance of the model's first-party API (e.g. OpenAI for o1) or the median across providers where a first-party API is not available (e.g. Meta's Llama models).