StarCoder2 and The Stack v2

In three linesHugging Face releases StarCoder2, open-source code model trained on The Stack v2 (17B code tokens dataset). StarCoder2 outperforms CodeLlama on HumanEval and MBPP benchmarks. Available in 3B, 7B, and 15B variants with open weights and source code.

Summary generated by Claude — human-verified