List your company

Reasoning on Reasoning Benchmarks Biology, Chemistry, Material, Physics, Tool Use

65.8Biology AccuracySRPO*
14.21627.6084154.392Jun 29, 2026

Evaluation results

MethodLinks
SRPO*Base Model=Qwen3-4B2026.0665.882.781.3746774.2
DRIFTBase Model=Qwen3-4B2026.0664.379.682.872.27474.6
DRIFTBase Model=Olmo3-7B2026.0661.38279.271.370.272.8
SC-SDPO*Base Model=Olmo3-7B2026.0655.981.579.171.162.970.1
GRPO*Base Model=Qwen3-4B2026.0655.578.380.171.962.969.7
DistIL*Base Model=Olmo3-7B2026.0655.38176.974.5——
SDPO*Base Model=Qwen3-4B2026.065477.374.366.761.166.7
SDPO*Base Model=Olmo3-7B2026.0652.88079.166.162.168
GRPO*Base Model=Olmo3-7B2026.0649.857.573.562.760.660.8
Qwen3-4BBase Model=Qwen3-4B, Configuration=Base model2026.0630.843.661.259.857.950.7
Olmo3-7BBase Model=Olmo3-7B, Configuration=Base model2026.0616.222.836.737.739.330.5