List your company

Clinical Task Execution on MedAgentBench OOD v2

87.1AccuracySkillX
25.63641.59357.5573.507May 28, 2026

Evaluation results

MethodLinks
SkillXModel=gpt-oss-120b2026.0587.1
Evo-MedAgentModel=gpt-oss-120b2026.0586.7
SkillXModel=GPT-5.4 (low)2026.0586.7
ExpeLModel=gpt-oss-120b2026.0586.3
ExpeLModel=GPT-5.4 (low)2026.0586.1
GRASPModel=gpt-oss-120b2026.0586
ExpeLModel=DeepSeek V4 Flash2026.0585
No skillsModel=GPT-4.12026.0584.9
ExpeLModel=GPT-4.12026.0584.4
Batch MemoryModel=GPT-4.12026.0583.9
No skillsModel=DeepSeek V4 Flash2026.0583.6
Seq. MemoryModel=GPT-4.12026.0583.3
Evo-MedAgentModel=GPT-4.12026.0582.8
GRASPModel=GPT-4.12026.0582.8
No skillsModel=GPT-5.4 (low)2026.0582.7
Seq. MemoryModel=gpt-oss-120b2026.0582.3
Seq. MemoryModel=GPT-5.4 (low)2026.0581.7
SkillXModel=GPT-4.12026.0580.6
GRASPModel=Gemini 3.1 Flash Lite2026.0580
Evo-MedAgentModel=GPT-5.4 (low)2026.0580
GRASPModel=GPT-5.4 (low)2026.0580
Batch MemoryModel=gpt-oss-120b2026.0578.8
Batch MemoryModel=GPT-5.4 (low)2026.0578.3
ExpeLModel=Gemini 3.1 Flash Lite2026.0577.2
No skillsModel=Gemini 3.1 Flash Lite2026.0576.3
No skillsModel=gpt-oss-120b2026.0575.2
Batch MemoryModel=Gemini 3.1 Flash Lite2026.0573.9
Evo-MedAgentModel=Gemini 3.1 Flash Lite2026.0572.2
GRASPModel=DeepSeek V4 Flash2026.0567
SkillXModel=Gemini 3.1 Flash Lite2026.0566.1
Seq. MemoryModel=DeepSeek V4 Flash2026.0565.3
Batch MemoryModel=DeepSeek V4 Flash2026.0557.3
Seq. MemoryModel=Gemini 3.1 Flash Lite2026.0556.1
Evo-MedAgentModel=DeepSeek V4 Flash2026.0542.7
SkillXModel=DeepSeek V4 Flash2026.0528