Clinical Benchmarks
Menu

Models

Every model with a sourced result here. The board count shows how widely it has been tested, not how well.

158 of 158 models

Index: the Clinical Benchmarks Index, this site’s own calculation from 10 boards, 0 to 100, printed with the number of boards it rests on. How it is computed

Models with results on this site
ModelLabKindReleasedIndex and what it rests onBoards with a resultBoards it appears on
Claude Opus 5AnthropicModel24 Jul 202663.26 of 10 boards8 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MAST (Medical AI Superintelligence Test), First, Do NOHARM (v2), PhysicianBench, HealthAgentBench, CHI-Bench
GPT-5.6 SolOpenAIModel9 Jul 202661.96 of 10 boards8 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), MAST (Medical AI Superintelligence Test), First, Do NOHARM (v2), HealthAgentBench, CHI-Bench
Claude Fable 5AnthropicModel9 Jun 202661.15 of 10 boards6 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), First, Do NOHARM (v2), CHI-Bench
Gemini 3.1 ProGoogleModel19 Feb 202658.09 of 10 boards10 / 13MedScribe (Vals AI), MedCode (Vals AI), MedPIC, MedXpertQA (MM), MAST (Medical AI Superintelligence Test), MedHELM, First, Do NOHARM (v2), PhysicianBench, EHR-Complex, HealthAdminBench
Muse SparkMetaModel8 Apr 202656.15 of 10 boards5 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), MedHELM
GPT-5OpenAIModel7 Aug 202554.15 of 10 boards5 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedPIC, First, Do NOHARM (v2)
GPT-5.2OpenAIModel11 Dec 202553.65 of 10 boards5 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedPIC, MedXpertQA (MM)
Claude Sonnet 5.5AnthropicModel28 Sep 202652.24 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
GPT-5.4OpenAIModel5 Mar 202651.87 of 10 boards10 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), MedHELM, PhysicianBench, EHR-Complex, HealthAgentBench, CHI-Bench, HealthAdminBench
Claude Opus 5.5AnthropicModel22 Sep 202651.84 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
GPT-5.5OpenAIModel23 Apr 202650.55 of 10 boards7 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), First, Do NOHARM (v2), PhysicianBench, HealthAgentBench, CHI-Bench
Claude Fable 5.1AnthropicModel1 Sep 202650.34 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
Claude Opus 4.8AnthropicModel28 May 202650.34 of 10 boards6 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), HealthAgentBench, CHI-Bench
Qwen3.5 397B A17BAlibabaModel16 Feb 202650.34 of 10 boards4 / 13MedXpertQA (MM), MAST (Medical AI Superintelligence Test), First, Do NOHARM (v2), EHR-Complex
Muse Spark 1.1MetaModel9 Jul 202650.33 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), First, Do NOHARM (v2)
Kimi K3Moonshot AIModel16 Jul 202650.24 of 10 boards5 / 13MedScribe (Vals AI), MedCode (Vals AI), MAST (Medical AI Superintelligence Test), First, Do NOHARM (v2), CHI-Bench
Gemini 2.5 ProGoogleModel17 Jun 202550.16 of 10 boards6 / 13MedScribe (Vals AI), MedCode (Vals AI), MedPIC, MedHELM, First, Do NOHARM (v2), EHR-Complex
Claude Sonnet 5AnthropicModel30 Jun 202647.35 of 10 boards6 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), MAST (Medical AI Superintelligence Test), PhysicianBench, CHI-Bench
Kimi K2.5Moonshot AIModel27 Jan 202645.54 of 10 boards5 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), EHR-Complex, HealthAdminBench
Claude Opus 4.6AnthropicModel5 Feb 202644.85 of 10 boards8 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), MedHELM, PhysicianBench, HealthAgentBench, CHI-Bench, HealthAdminBench
GPT-6 AstraOpenAIModel3 Sep 202644.03 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
Gemini 3.5 FlashGoogleModel19 May 202643.83 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), MedHELM
Gemini 3 ProGoogleModel18 Nov 202543.02 of 10 boards2 / 13MedScribe (Vals AI), MedXpertQA (MM)
GPT-6.1 SolOpenAIModel29 Sep 202642.33 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
Gemini 3.6 FlashGoogleModel21 Jul 202641.83 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), MAST (Medical AI Superintelligence Test)
Claude Opus 4.5AnthropicModel24 Nov 202541.63 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM)
Qwen3.8 MaxAlibabaModel3 Aug 202641.63 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM)
Grok 4.7SpaceX AIModel21 Sep 202640.73 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
GPT-6 SolOpenAIModel22 Sep 202640.33 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
Claude Opus 4.7AnthropicModel16 Apr 202640.14 of 10 boards6 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench, HealthAgentBench, CHI-Bench
GPT-6 LunaOpenAIModel22 Sep 202639.73 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
GPT-5.6 TerraOpenAIModel9 Jul 202638.33 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), CHI-Bench
GPT-5.6 LunaOpenAIModel9 Jul 202637.53 of 10 boards4 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), CHI-Bench
GPT-5.1OpenAIModel12 Nov 202536.93 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
Grok 4.6SpaceX AIModel12 Aug 202636.63 of 10 boards3 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI)
GLM 5.3 FlashZhipuModel26 Aug 202636.31 of 10 boards1 / 13MedScribe (Vals AI)
DeepSeek V4 ProDeepSeekModel24 Apr 202636.24 of 10 boards5 / 13MedScribe (Vals AI), MedCode (Vals AI), MedPIC, PhysicianBench, CHI-Bench
GLM 5.1ZhipuModel7 Apr 202636.23 of 10 boards4 / 13MedScribe (Vals AI), MedCode (Vals AI), First, Do NOHARM (v2), CHI-Bench
Grok 4.3SpaceX AIModelNot stated35.03 of 10 boards4 / 13MedScribe (Vals AI), MedCode (Vals AI), MAST (Medical AI Superintelligence Test), CHI-Bench
Gemini 3.7 FlashGoogleModel13 Aug 202634.62 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Gemini 3 FlashGoogleModel17 Dec 202534.02 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), CHI-Bench
Gemini 2.5 FlashGoogleModel17 Jun 202533.91 of 10 boards1 / 13MedScribe (Vals AI)
Muse Spark 1.2MetaModel5 Aug 202633.62 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Kimi K2.6Moonshot AIModel20 Apr 202633.04 of 10 boards5 / 13MedScribe (Vals AI), MedCode (Vals AI), First, Do NOHARM (v2), PhysicianBench, CHI-Bench
DeepSeek V4 FlashDeepSeekModel24 Apr 202632.81 of 10 boards1 / 13MedScribe (Vals AI)
Qwen3.6 PlusAlibabaModel1 Apr 202632.34 of 10 boards4 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), PhysicianBench
Gemini 3.8 FlashGoogleModel2 Sep 202632.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude Sonnet 4.6AnthropicModel17 Feb 202632.14 of 10 boards6 / 13HealthBench Professional, MedPIC, PhysicianBench, EHR-Complex, HealthAgentBench, CHI-Bench
MiniMax M3MiniMaxModel1 Jun 202631.82 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
MiMo V2.6 ProXiaomiModel22 Sep 202631.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude Haiku 5.5AnthropicModel7 Oct 202630.92 of 10 boards2 / 13HealthBench Professional, PhysicianBench
o3OpenAIModel16 Apr 202530.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude Opus 4.1AnthropicModel5 Aug 202530.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude Sonnet 4.5AnthropicModel29 Sep 202530.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
GLM 5.3ZhipuModel18 Aug 202630.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Grok 4.5SpaceX AIModel8 Jul 202630.42 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Hy4 PreviewTencentModel28 Aug 202630.02 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
DeepSeek R1DeepSeekModel20 Jan 202529.52 of 10 boards2 / 13MedHELM, First, Do NOHARM (v2)
Gemini 3.1 Flash Lite PreviewGoogleModel3 Mar 202629.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
GPT-5 miniOpenAIModel7 Aug 202529.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3.5-PlusAlibabaModelNot stated29.51 of 10 boards1 / 13MedPIC
DeepSeek V4.1 FlashDeepSeekModel10 Sep 202629.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
InklingThinking MachinesModel15 Jul 202629.32 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), CHI-Bench
MiMo V2.6 FlashXiaomiModel22 Sep 202629.22 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3.7 PlusAlibabaModel31 May 202629.01 of 10 boards1 / 13MedXpertQA (MM)
GLM 5.2ZhipuModel16 Jun 202628.92 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), CHI-Bench
GPT OSS 120BOpenAIModel5 Aug 202528.81 of 10 boards1 / 13MedPIC
Gemini 3.5 Flash LiteGoogleModel21 Jul 202628.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3.5-27BAlibabaModelNot stated28.31 of 10 boards1 / 13MedPIC
GPT-5.4 nanoOpenAIModel17 Mar 202628.22 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3.5-35B-A3BAlibabaModelNot stated27.91 of 10 boards1 / 13MedPIC
Inkling SmallThinking MachinesModel30 Jul 202627.62 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen 3.7 MaxAlibabaModelNot stated27.42 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Grok 4 Fast (Reasoning)SpaceX AIModel19 Sep 202527.12 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Grok 4.20SpaceX AIModel10 Mar 202627.04 of 10 boards4 / 13MedScribe (Vals AI), MedCode (Vals AI), MedXpertQA (MM), PhysicianBench
Grok 4SpaceX AIModel9 Jul 202526.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3.5-9BAlibabaModelNot stated26.01 of 10 boards1 / 13MedPIC
MiniMax-M2.1MiniMaxModelNot stated25.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Gemma 4 31BGoogleModel2 Apr 202625.01 of 10 boards1 / 13MedXpertQA (MM)
MedGemma 27B TextGoogleModelNot stated25.01 of 10 boards1 / 13MedPIC
Claude Sonnet 4AnthropicModel22 May 202524.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Ling 3.0 FlashAnt GroupModelNot stated24.72 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Fleming-R1-32BUbiquantModelNot stated24.71 of 10 boards1 / 13MedPIC
Lingshu-32BAlibabaModelNot stated24.31 of 10 boards1 / 13MedPIC
DeepSeek-V3.2-ExpDeepSeekModel29 Sep 202524.11 of 10 boards1 / 13EHR-Complex
GPT OSS 20BOpenAIModel5 Aug 202524.11 of 10 boards1 / 13MedPIC
o4-miniOpenAIModel16 Apr 202523.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen 3 VL PlusAlibabaModelNot stated23.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Mistral Medium 3.5MistralModel28 Apr 202623.72 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen 3.5 FlashAlibabaModelNot stated23.72 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Gemma 4 26B A4BGoogleModel2 Apr 202623.71 of 10 boards1 / 13MedXpertQA (MM)
Grok 4 Fast (Non-Reasoning)SpaceX AIModel19 Sep 202523.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
GLM 4.7ZhipuModel22 Dec 202523.42 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
MiMo V2.5XiaomiModelNot stated23.42 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen 3.8 27BAlibabaModelNot stated23.42 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Llama 4 MaverickMetaModel5 Apr 202523.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen 3 Max ThinkingAlibabaModelNot stated23.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude Haiku 4.5AnthropicModel15 Oct 202523.14 of 10 boards5 / 13HealthBench Professional, MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench, CHI-Bench
MiMo V2.5 ProXiaomiModelNot stated23.13 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
GPT-5 nanoOpenAIModel7 Aug 202522.82 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
DeepSeek-V3.1DeepSeekModel21 Aug 202522.81 of 10 boards1 / 13EHR-Complex
Ling 3.0 Flash FinAnt GroupModelNot stated22.72 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Llama 3.1 70BMetaModel23 Jul 202422.71 of 10 boards1 / 13MedPIC
Grok 4.1 Fast (Reasoning)SpaceX AIModel19 Nov 202522.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Grok 4.1 Fast Non-ReasoningSpaceX AIModel19 Nov 202522.52 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
GPT-5.4 miniOpenAIModel17 Mar 202622.51 of 10 boards3 / 13MedHELM, HealthAgentBench, CHI-Bench
HuatuoGPT-o1-70BFreedomIntelligenceModelNot stated22.51 of 10 boards1 / 13MedPIC
Qwen3-32B-SFTAnt GroupModelNot stated22.41 of 10 boards1 / 13EHR-Complex
Gemini 2.5 Flash LiteGoogleModel22 Jul 202521.72 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Qwen3-235B-A22B-Instruct-2507AlibabaModelNot stated21.61 of 10 boards1 / 13EHR-Complex
Gemini 3 Pro (11/25)GoogleModelNot stated21.31 of 10 boards1 / 13MedCode (Vals AI)
MiniMax-M2.7MiniMaxModel18 Mar 202621.13 of 10 boards3 / 13MedScribe (Vals AI), MedCode (Vals AI), PhysicianBench
Mercury 2.5InceptionModel8 Sep 202621.02 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Fleming-R1-7BUbiquantModelNot stated20.71 of 10 boards1 / 13MedPIC
GPT-4.1 miniOpenAIModel14 Apr 202520.01 of 10 boards1 / 13EHR-Complex
Gemma 4 12BGoogleModel3 Jun 202619.91 of 10 boards1 / 13MedXpertQA (MM)
Baichuan-M2-32BBaichuanModelNot stated19.51 of 10 boards1 / 13MedPIC
Qwen3-VL-235B-A22BAlibabaModelNot stated19.41 of 10 boards1 / 13MedXpertQA (MM)
GPT-4.1OpenAIModel14 Apr 202519.21 of 10 boards1 / 13EHR-Complex
Laguna M.1PoolsideModel28 Apr 202618.62 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Claude 3.7 SonnetAnthropicModel24 Feb 202518.41 of 10 boards1 / 13MedHELM
Qwen3-14B-SFTAnt GroupModelNot stated18.41 of 10 boards1 / 13EHR-Complex
HuatuoGPT-o1-8BFreedomIntelligenceModelNot stated18.11 of 10 boards1 / 13MedPIC
Lingshu-7BAlibabaModelNot stated17.71 of 10 boards1 / 13MedPIC
HealthGPT-Pro-8BZJU4HealthCareModelNot stated17.41 of 10 boards1 / 13MedPIC
Laguna XS.2PoolsideModel28 Apr 202617.32 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Hulu-Med-7BZJU-AI4HModel8 Oct 202517.11 of 10 boards1 / 13MedPIC
DeepSeek V4 Flash 0731DeepSeekModel31 Jul 202616.91 of 10 boards1 / 13MedCode (Vals AI)
Llama 4 ScoutMetaModel5 Apr 202516.82 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
Gemini 2.5 Flash (7/17)GoogleModelNot stated16.51 of 10 boards1 / 13MedCode (Vals AI)
Gemini 2.5 Flash Preview (9/25)GoogleModel25 Sep 202516.51 of 10 boards1 / 13MedCode (Vals AI)
Gemma 3 27BGoogleModel12 Mar 202516.41 of 10 boards1 / 13MedPIC
Llama 3.1 8BMetaModel23 Jul 202416.31 of 10 boards1 / 13MedPIC
Gemma 3 12BGoogleModel12 Mar 202516.21 of 10 boards1 / 13MedPIC
MedGemma 4BGoogleModelNot stated16.21 of 10 boards1 / 13MedPIC
Command A+CohereModel20 May 202615.92 of 10 boards2 / 13MedScribe (Vals AI), MedCode (Vals AI)
GPT-5.5 InstantOpenAIModel5 May 202615.71 of 10 boards1 / 13HealthBench Professional
Nemotron 3 UltraNVIDIAModel4 Jun 202615.71 of 10 boards1 / 13MedCode (Vals AI)
HealthGPT-Pro-4BZJU4HealthCareModelNot stated14.71 of 10 boards1 / 13MedPIC
Qwen3-32BAlibabaModel29 Apr 202514.71 of 10 boards1 / 13EHR-Complex
MAI-Thinking-1MicrosoftModel12 Aug 202614.31 of 10 boards1 / 13HealthBench Professional
Gemini 2.0 FlashGoogleModel5 Feb 202514.01 of 10 boards1 / 13MedHELM
Gemini 2.5 Flash Lite (9/25)GoogleModel25 Sep 202514.01 of 10 boards1 / 13MedCode (Vals AI)
GPT-4oOpenAIModel13 May 202412.61 of 10 boards1 / 13EHR-Complex
Qwen3-14BAlibabaModel29 Apr 202512.21 of 10 boards1 / 13EHR-Complex
Gemma 4 E4BGoogleModel2 Apr 202611.71 of 10 boards1 / 13MedXpertQA (MM)
Gemma 4 E2BGoogleModel2 Apr 20269.61 of 10 boards1 / 13MedXpertQA (MM)
Qwen3-4BAlibabaModel29 Apr 20256.51 of 10 boards1 / 13EHR-Complex
Nemotron 3.5 LightningNVIDIAModel11 Aug 20261.81 of 10 boards1 / 13MedScribe (Vals AI)
Doximity Ask 6.1DoximityProductNot statedNot ranked1 / 13First, Do NOHARM (v2)
Glass 5.6 MaxGlass HealthProductNot statedNot ranked1 / 13First, Do NOHARM (v2)
GPT-5.3-CodexOpenAIModel5 Feb 2026Not ranked1 / 13HealthAgentBench
LiSA 2.5AMBOSSProductNot statedNot ranked1 / 13First, Do NOHARM (v2)
MedGuardMedGuardProductNot statedNot ranked1 / 13CHI-Bench
NVIDIA Nemotron 3 Ultra 550B A55BNVIDIAModel4 Jun 2026Not ranked1 / 13CHI-Bench
OpenEvidenceOpenEvidenceProductNot statedNot ranked1 / 13First, Do NOHARM (v2)
Qwen 3.5AlibabaModelNot statedNot ranked1 / 13HealthAdminBench
Qwen3.6-MaxAlibabaModelNot statedNot ranked1 / 13CHI-Bench