List your company

Medical Calculation on MedCalc-Bench

81.4AccuracyGPT-5
4.33624.34344.3564.357May 10, 2026May 18, 2026May 27, 2026Jun 5, 2026Jun 13, 2026Jun 22, 2026Jul 1, 2026

Evaluation results

MethodLinks
GPT-5Family=Frontier API2026.0581.4——————
Claude Sonnet 4.5Backbone=Claude Sonnet 4.52026.0776——————
DeepSeek-R1Backbone=DeepSeek-R12026.0772.7——————
GPT-4.1Family=Frontier API2026.0572——————
GPT-4.1-miniFamily=Frontier API2026.0567.5——————
FaithMed (Qwen3 4B)Backbone=Qwen3 4B2026.0756.5——————
MedGemma-27BFamily=Medical / open-source2026.0551.9——————
MedGemma 27BBackbone=MedGemma 27B2026.0749——————
CLR-voyance-8BFamily=8B family, Merge Strategy=DELLA-Linear2026.0546——————
Qwen3 8BBackbone=Qwen3 8B2026.0744.9——————
Qwen3-8B baseFamily=8B family2026.0541.5——————
Agentic Search + GRPO (Qwen3 4B)Backbone=Qwen3 4B2026.0739——————
FaithMed (Qwen3 1.7B)Backbone=Qwen3 1.7B2026.0733.2——————
DS-R1-0528Family=Medical / open-source2026.0531.6——————
Agentic Search (Qwen3 4B)Backbone=Qwen3 4B2026.0729.6——————
Agentic Search + GRPO (Qwen3 1.7B)Backbone=Qwen3 1.7B2026.0727.1——————
HuatuoGPT-o1Family=Medical / open-source2026.0527——————
Gemma 3 12BBackbone=Gemma 3 12B2026.0727——————
Llama 3.1 8B InstructBackbone=Llama 3.1 8B2026.0721.4——————
MG-BFamily=4B family, Merge Strategy=Breadcrumbs2026.0521——————
MG-B+AFamily=4B family, Merge Strategy=Breadcrumbs+AIM2026.0518.5——————
Base (Qwen3 1.7B)Backbone=Qwen3 1.7B2026.0716.9——————
MG-Δ+AFamily=4B family, Merge Strategy=DELLA-Linear+AIM2026.0515.5——————
Agentic Search (Qwen3 1.7B)Backbone=Qwen3 1.7B2026.0715.3——————
CLR-voyance-4BFamily=4B family, Merge Strategy=DELLA-Linear2026.0515——————
Base (Qwen3 4B)Backbone=Qwen3 4B2026.0713.7——————
MG-4B baseFamily=4B family2026.0510.7——————
MedRAG (Qwen3 4B)Backbone=Qwen3 4B2026.0710.3——————
DS-R1-DistillFamily=Medical / open-source2026.058.6——————
MedRAG (Qwen3 1.7B)Backbone=Qwen3 1.7B2026.078.4——————
Gemma-3-4BFamily=Medical / open-source2026.057.3——————
Default (Greedy) CoTMetric=Exact-Match2026.01—0.140.240.430.380.410.32
Default (Sampled) CoTMetric=Exact-Match2026.01—0.160.290.450.380.350.33
Empty CoTMetric=Exact-Match2026.01—0.10.180.450.360.380.29
Ens. (QwQ+DAPO/OSS)Metric=Exact-Match2026.01—0.390.370.410.410.40.4
Ens. (QwQ+OpenT/OSS)Metric=Exact-Match2026.01—0.340.410.420.380.40.39
Ens. (QwQ+OSS/DAPO)Metric=Exact-Match2026.01—0.280.370.450.420.430.38
Trans. CoT (DAPO)Metric=Exact-Match2026.01—0.310.40.390.40.410.38
Trans. CoT (NRR)Metric=Exact-Match2026.01—0.140.150.240.30.340.23
Trans. CoT (OpenT)Metric=Exact-Match2026.01—0.210.240.260.260.250.24
Trans. CoT (OSS)Metric=Exact-Match2026.01—0.260.440.430.40.440.39
Trans. CoT (QwQ)Metric=Exact-Match2026.01—0.340.370.390.380.370.37