Back to directory
Bigcodebench
Free502[ICLR'25] BigCodeBench: Benchmarking Code Generation Towards AGI
About
BigCodeBench measures how well AI models generate code by testing them against a large set of programming tasks. It's built for researchers evaluating code generation systems. The benchmark includes over 1,000 problems spanning different difficulty levels and programming domains.
Key Features
- agent
- agents
- benchmark
- chatgpt
- claude-3
Pricing
FreeOpen source. You supply your own LLM API keys.
Categories
CodingResearch