List your company

Clinical Task Execution on MedAgentBench (val)

86.2AccuracyGRASP
26.9242.3157.773.09May 28, 2026

Evaluation results

MethodLinks
GRASPModel=GPT-4.12026.0586.2
GRASPModel=gpt-oss-120b2026.0586
GRASPModel=GPT-5.4 (low)2026.0584.6
GRASPModel=DeepSeek V4 Flash2026.0568.8
GRASPModel=Gemini 3.1 Flash Lite2026.0566.7
Evo-MedAgentModel=gpt-oss-120b2026.0562.5
SkillXModel=DeepSeek V4 Flash2026.0555
SkillXModel=Gemini 3.1 Flash Lite2026.0554.2
Evo-MedAgentModel=DeepSeek V4 Flash2026.0552.5
SkillXModel=gpt-oss-120b2026.0551.5
ExpeLModel=Gemini 3.1 Flash Lite2026.0551.2
ExpeLModel=DeepSeek V4 Flash2026.0550.7
Evo-MedAgentModel=Gemini 3.1 Flash Lite2026.0549.6
No skillsModel=Gemini 3.1 Flash Lite2026.0549
Seq. MemoryModel=Gemini 3.1 Flash Lite2026.0548.8
Batch MemoryModel=GPT-5.4 (low)2026.0547.9
ExpeLModel=gpt-oss-120b2026.0547.5
Seq. MemoryModel=GPT-5.4 (low)2026.0546.7
Batch MemoryModel=GPT-4.12026.0546.2
SkillXModel=GPT-5.4 (low)2026.0545.4
ExpeLModel=GPT-5.4 (low)2026.0545
Seq. MemoryModel=GPT-4.12026.0542.5
ExpeLModel=GPT-4.12026.0542.5
Evo-MedAgentModel=GPT-4.12026.0542.5
Evo-MedAgentModel=GPT-5.4 (low)2026.0542.5
No skillsModel=DeepSeek V4 Flash2026.0541.8
No skillsModel=GPT-5.4 (low)2026.0540.6
No skillsModel=GPT-4.12026.0540.1
Seq. MemoryModel=gpt-oss-120b2026.0538.5
Batch MemoryModel=gpt-oss-120b2026.0536.8
Batch MemoryModel=DeepSeek V4 Flash2026.0536.5
No skillsModel=gpt-oss-120b2026.0536.4
SkillXModel=GPT-4.12026.0533.8
Seq. MemoryModel=DeepSeek V4 Flash2026.0533
Batch MemoryModel=Gemini 3.1 Flash Lite2026.0529.2