BigCodeBench

1,140 Python tasks that require chaining calls across 139 real libraries, testing whether a model can use diverse tools correctly rather than write self-contained algorithmic code.

unassessed

This page is a discovery lead. Nobody has yet assessed it against the catalogue contract, so it carries no disposition. Absence of evidence here is not evidence of staleness.
Categorycoding
Subcategoryfunction-calling / library-use code generation
Page statusactive
MetricPass@1 (including a "calibrated" variant that accounts for omitted setup code)
Directionhigher_is_better
Unit%
Dataset size1140
Dataset licenceApache-2.0
PublisherBigCode project (multi-institution community collaboration)

What it measures

BigCodeBench asks a model to write a Python function that correctly uses one or more calls into real, popular libraries (data processing, visualization, networking, cryptography and more) to satisfy either a structured docstring (the Complete split) or a condensed natural-language instruction (the Instruct split). This targets a different skill than algorithmic benchmarks like HumanEval or MBPP: knowing which library function to call, with which arguments, and how to combine several such calls correctly, rather than implementing logic from scratch in the standard library alone.

Task format

Given a function signature with a docstring (Complete) or a natural-language instruction (Instruct), the model generates a Python function body; the generated code is executed against an average of 5.6 unit test cases per task with about 99% branch coverage.

Models reporting this benchmark

No model card in ModelSpec reports this benchmark yet.

Data

This page as JSON · Edit on GitHub