arxiv
Qwen3-Coder + CodeStruct
1 benchmarks
SWE-bench Verified
#45 of 72
66.2%
benchmark.space canonical public ledger of AI capability