DeepSeek-V3.2 Technical Report
Open paper
Benchmark
Model
Rank
Results
code-generation-on-humaneval
DeepSeek V3.2 Thinking
#10
Pass@1: 93.9
code-generation-on-swe-bench-pro
deepseek-v3p2
#18
Resolved: 15.56
Rank counts only results with a code link.