Agent Evaluation and Benchmarks🌱 growing
AgentBenchTHUDM/AgentBench
Comprehensive benchmark for evaluating LLMs as agents across 8 distinct environments.
Python#Benchmark
GitHub ↗← Back to ExploreGitHub ↗
Agent Evaluation and Benchmarks • 🚀 production
SWE-bench
SWE-bench/SWE-benchBenchmark for evaluating LLMs on real-world software engineering tasks from GitHub issues.
#Python#GitHub
Related Resources in Agent Evaluation and Benchmarks
Agent Evaluation and Benchmarks🌱 growing
Frontier benchmark for measuring general intelligence capabilities in AI agents beyond pattern matching.
Python#Benchmark
Website ↗Agent Evaluation and Benchmarks🔬 emerging
ClawBenchTIGER-AI-Lab/ClawBench
Evaluates web agents on 283 real-world tasks across 163 live websites with interception and trace-based scoring.
Python#Benchmark
GitHub ↗