List your company

Multiple-Choice Question Answering on PubHealthBench

99.5Blood Safety & STIs/HIV AccuracyLlama-3.3-70B
63.72473.01282.391.588Jul 7, 2026

Evaluation results

MethodLinks
Llama-3.3-70Bmodel used to generate benchmark=true, number of parameters=70B2026.0799.599.710099.410099.899.210099.299.799.5
MedGemma-27Bnumber of parameters=27B2026.0799.599.510098.899.299.699.698.39999.399.2
Gemma-3-27Bnumber of parameters=27B2026.0799.299.210098.899.299.699.298.79998.898.9
Command-R-32Bnumber of parameters=32B2026.0798.998.899.797.999.299.298.397.498.699.198.7
Gemma-3-12Bnumber of parameters=12B2026.0798.999.499.798.598.499.399.299.698.599.298.9
Phi-4-14Bnumber of parameters=14B2026.0798.699.210098.598.699.297.999.199.19998.9
Gemma-2-27Bnumber of parameters=27B2026.0798.498.899.798.898.69997.597.498.498.798.5
Llama-3.1-8Bnumber of parameters=8B2026.0797.69899.797.997.898.997.997.997.698.297.9
Gemma-3-4Bnumber of parameters=4B2026.0795.7979896.295.396.596.294.995.895.695.6
Phi-4-4Bnumber of parameters=4B2026.0795.496.998.795.9979895.897.997.597.397.1
Gemma-3-1Bnumber of parameters=1B2026.0765.169.264.764.961.967.973.263.465.865.666.4