搜索 | EvalScope

Skip to content

文档博客

⌘ K

🚀 快速开始

简介
安装
快速上手
可视化
参数说明
支持的数据集
- LLM评测集
  - AA-LCR
  - AIME-2024
  - AIME-2025
  - AIME-2026
  - AlpacaEval2.0
  - AMC
  - AnatEM
  - ARC
  - ArenaHard
  - ArxivRollBench
  - ArxivRollBench-Full
  - BBH
  - BC2GM
  - BC4CHEMD
  - BC5CDR
  - BioMixQA
  - BroadTwitterCorpus
  - C-Eval
  - Chinese-SimpleQA
  - CL-bench
  - C-MMLU
  - CoinFlip
  - CommonsenseQA
  - Competition-MATH
  - CoNLL2003
  - CoNLL++
  - Copious
  - CrossNER
  - Data-Collection
  - DocMath
  - DrivelologyBinaryClassification
  - DrivelologyMultilabelClassification
  - DrivelologyNarrativeSelection
  - DrivelologyNarrativeWriting
  - DROP
  - EQ-Bench
  - FinNER
  - FRAMES
  - GeneralArena
  - General-MCQ
  - General-QA
  - GeniaNER
  - GPQA-Diamond
  - GSM8K
  - HaluEval
  - HarveyNER
  - HealthBench
  - HellaSwag
  - Humanity's-Last-Exam
  - HMMT25
  - HumanEval
  - HumanEvalPlus
  - IFBench
  - IFEval
  - IQuiz
  - JNLPBA
  - JNLPBA-Rare
  - Live-Code-Bench
  - LogiQA
  - LongBench-v2
  - MaritimeBench
  - MATH-500
  - MathQA
  - MBPP
  - MBPP-Plus
  - Med-MCQA
  - MGSM
  - Minerva-Math
  - MIT-Movie-Trivia
  - MIT-Restaurant
  - MMLU
  - MMLU-Pro
  - MMLU-Redux
  - MMMLU
  - MRI-MCQA
  - Multi-IF
  - MultiNERD
  - MultiPL-E HumanEval
  - MultiPL-E MBPP
  - MusicTrivia
  - MuSR
  - NCBI
  - Needle-in-a-Haystack
  - OntoNotes5
  - OpenAI MRCR
  - PIQA
  - PolyMath
  - ProcessBench
  - PubMedQA
  - QASC
  - RACE
  - RefCOCO
  - SciCode
  - SciQ
  - SimpleQA
  - SIQA
  - SuperGPQA
  - SWE-bench_Lite
  - SWE-bench_Verified
  - SWE-bench_Verified_mini
  - ToolBench-Static
  - TriviaQA
  - TruthfulQA
  - TweeBankNER
  - TweetNER7
  - Winogrande
  - WMT2024++
  - WNUT2017
  - ZebraLogicBench
- VLM评测集
  - A-OKVQA
  - AI2D
  - AIR-Bench-Chat
  - AIR-Bench-Foundation
  - BLINK
  - CCBench
  - ChartQA
  - CMMMU
  - CMMU
  - DocVQA
  - FLEURS
  - General-VMCQ
  - General-VQA
  - GSM8K-V
  - HallusionBench
  - InfoVQA
  - LibriSpeech
  - MathVerse
  - MathVision
  - MathVista
  - MIA-Bench
  - MicroVQA
  - MMBench
  - MMStar
  - MMMU
  - MMMU-PRO
  - MVBench
  - OCRBench
  - OCRBench-v2
  - OlympiadBench
  - OmniBench
  - OmniDocBench
  - POPE
  - RealWorldQA
  - ScienceQA
  - SEED-Bench-2-Plus
  - SimpleVQA
  - TIR-Bench
  - TORGO
  - Video-MME-v2
  - VisuLogic
  - V*Bench
  - ZeroBench
- AGENT评测集
- AIGC评测集
  - EvalMuse
  - GEdit-Bench
  - GenAI-Bench
  - general_t2i
  - HPD-v2
  - TIFA-160
- 其他数据集
❓ 常见问题

🔧 教程

其他评测后端
- OpenCompass
- VLMEvalKit
- RAGEval
模型推理性能压测
AIGC 评测
- 文生图评测
- 图片编辑评测
竞技场模式
沙箱环境使用
Agent 评测
- 内置 AgentLoop 模式
- 外部 Agent Bridge 模式
EvalScope 服务部署

🛠️ 进阶教程

构建评测指数（Index）
自定义评测数据集
自定义模型评测
👍 贡献基准评测

🧰 扩展评测基准

扩展评测基准
- Terminal-Bench
- GAIA
- SWE-bench
- SWE-bench_Pro
- τ-bench
- τ²-bench
- τ³-bench
- BFCL-v3
- BFCL-v4
- 大海捞针测试
- ToolBench
- LongBench-Write

📖 最佳实践

最佳实践

🧪 基准测试结果

基准测试
- MMLU
速度基准测试
- QwQ-32B-Preview

🌟 博客

欢迎阅读 EvalScope 博客！

/

搜索

Please activate Javascript to enable searching the documentation.

© 2022-2025, Alibaba ModelScope Built with Sphinx 9.1.0