Scientific Code Benchmark
SciCode evaluates language models on generating code for realistic scientific research problems across 16 subfields of physics, math, chemistry, biology, and material science. Problems decompose into 338 subproblems requiring domain knowledge recall, scientific reasoning, and precise code synthesis. Based on real scripts from published research.
ModelsModels with a result.35
Top-3 spreadPoints from first to third.1.4 pts
YearYear of release.2024
Result and price
17 of 35 results
Results
35 results#ModelResultThe score from the source.
