SOTA Clinical Reasoning on MIMIC-CDM-FI and PapersWithCode

90.1Accuracy

DeepSeek-V4

Updated 1mo ago

Evaluation Results

Method	Links
DeepSeek-V4 2026.05		90.1
MedGuideX-9B 2026.05		85.2
GPT-5.0 2026.05		85.1
RL with CPG-Derived Process Rewards 2026.05		83.4
Qwen3.5-9B 2026.05		82.5
MedGuideX-4B 2026.05		82.4
Qwen3.5-4B 2026.05		81.9
Claude-Haiku-4.5 2026.05		81.6
Qwen3.5-9B 2026.05		81.6
Qwen3.5-4B 2026.05		81.4
Llama-Aloe-Beta-8B 2026.05		80.8
Qwen3.5-4B 2026.05		80.4
MedGemma-1.5-4B 2026.05		80.1
Qwen3.5-9B 2026.05		80.1
Hulu-Med-7B 2026.05		78.7
Fine-tuning with CPG 2026.05		78.4
Lingshu-7B 2026.05		78
Kimi-2.6 2026.05		74.6
MedReason-8B 2026.05		73
MediPhi-Instruct-4B 2026.05		72.6
HuatuoGPT-o1-8B 2026.05		68.1
CPGPrompt 2026.05		67.6
Clinical-R1-3B 2026.05		62.3
Llava-Med-v1.5-7B 2026.05		60.1
MedAlpaca-7B 2026.05		41
BioMistral-7B 2026.05		35.6