List your company

Code Generation on MBPP CodaSet OOD (test)

79.37Performance (%)Qwen3-235B
71.663673.664375.66577.6657May 25, 2026

Evaluation results

MethodLinks
Qwen3-235B2026.0579.3713.8
Kimi-K2-Instruct2026.0577.7820.1
DecoRBase Model=DeepSeek-V3.12026.0576.7212.4
gemma-3-27b-it2026.0576.191.3
gemma-3-12b-it2026.0576.191
DeepSeek-V3.12026.0575.6614.1
DeepSeek-V3.2-Exp2026.0575.1312.2
MODEL-SAT2026.0574.64.7
EmbedLLM2026.0573.728.4
gpt-oss-120b2026.0573.5411.6
KNN Router2026.0573.542.7
LLM Router2026.0573.194
Mistral-3.2-24B2026.0573.026
RouterDC2026.0572.666.9
DecoRBase Model=gpt-oss-120b2026.0572.4910.2
Random Router2026.0571.968.9