Clinical Benchmarks Index
Claude Sonnet 5.5 is first at 63.2. Claude Opus 5.5 follows at 62.8, then Claude Opus 5 at 56.8.
This site's own calculation from 12 of 14 clinical boards, each read on its own 0 to 100 scale. Method of 31 October 2026.
- Models ranked
- 128
- Results
- 472
- Source documents
- 92
Ranking, model by model
The marks under each column show which of the 12 index boards the score uses.
Select a column to open the model. The filter hides models and leaves every rank as it is.
Each index board, model by model
The top 10 models on each of the 12 index boards, on that board's own 0 to 100 scale. Each column opens its source.
HealthBench Professional
- 0.703GPT-6 Astra
- 0.692Claude Sonnet 5.5
- 0.660Claude Fable 5
- 0.656Claude Opus 5.5
- 0.648Claude Haiku 5.5
- 0.642GPT-6.1 Sol
- 0.627GPT-6.1 Terra
- 0.623Muse Wonder Full
- 0.621Claude Fable 5.1
- 0.608GPT-6 Luna
37 modelsFull board
MedScribe (Vals AI)near ceiling
- 91.43Claude Opus 5.5
- 91.29Claude Fable 5.1
- 91.10Claude Sonnet 5.5
- 90.98Claude Opus 5
- 90.06Muse Spark 1.2
- 89.38Grok 4.7
- 88.89Muse Spark 1.1
- 88.94GLM 5.3 Flash
- 88.81GLM 5.3
- 88.52Claude Fable 5
95 modelsFull board
MedCode (Vals AI)
- 63.57Claude Opus 5
- 59.06Gemini 3.1 Pro
- 58.80Gemini 4 Argon
- 56.07Claude Fable 5
- 55.92Gemini 3 Flash
- 55.83Gemini 3.5 Flash
- 54.86Claude Opus 4.7
- 53.51Claude Fable 5.1
- 53.39Gemini 3.7 Flash
- 53.22Claude Opus 4.8
95 modelsFull board
MedXpertQA (MM)near ceiling
- 81.5GPT-5.6 Sol
- 81.3Gemini 3.1 Pro
- 80.4Qwen3.8 Max
- 80.0Claude Fable 5
- 78.4Muse Spark
- 77.1GPT-5.4
- 76.0Gemini 3 Pro
- 73.3GPT-5.2
- 71.7Claude Opus 4.8
- 71.0Qwen3.7 Plus
22 modelsFull board
MAST (Medical AI Superintelligence Test)
- 60.2GPT-5.6 Sol
- 60.1Kimi K3
- 59.3Gemini 3.6 Flash
- 58.9Gemini 3.1 Pro
- 57.9Qwen3.5 397B A17B
- 57.1Claude Opus 5
- 56.6Claude Sonnet 5
- 53.7Grok 4.3
8 modelsFull board
First, Do NOHARM (v2)
- 79.7Muse Spark 1.1
- 74.6Claude Opus 5
- 74.0Kimi K3
- 70.1GPT-5.6 Sol
- 70.0GPT-5.5
- 68.6GPT-5
- 65.0Claude Fable 5
- 62.6Gemini 3.1 Pro
- 61.9Gemini 2.5 Pro
- 61.1Qwen3.5 397B A17B
13 modelsFull board
PhysicianBench
- 68.4Claude Opus 5.5
- 63.2Claude Sonnet 5.5
- 61.0Claude Fable 5.1
- 57.6Claude Opus 5
- 46.3GPT-5.5
- 43.0Claude Haiku 5.5
- 37.4Claude Sonnet 5
- 31.7Claude Opus 4.6
- 29.3Claude Opus 4.7
- 27.7GPT-5.4
19 modelsFull board
Bedside Benchnear ceiling
- 94.6Claude Opus 5
- 94.3Muse Spark Open
- 93.4GPT-6 Astra
- 92.7Gemini 4 Argon
- 91.1GPT-5.6 Sol
- 89.7Kimi K3
- 89.1Ember-1
- 88.1GPT-6 Sol
- 87.2GPT-6 Luna
- 86.9GLM 5.3
15 modelsFull board
RCM-Benchnear ceiling
- 92.68Claude Opus 5
- 91.49Gemini 4 Argon
- 91.21GPT-5.6 Sol
- 91.24Claude Sonnet 5
- 91.05Claude Sonnet 5.5
- 90.23Gemini 3.8 Flash
- 90.14Claude Opus 5.5
- 89.82Claude Sonnet 4.6
- 88.99DeepSeek V4.1 Flash
- 88.38Claude Haiku 5.5
20 modelsFull board
DAYJOB Health
- 48.3Claude Opus 5.5
- 34Claude Sonnet 5.5
- 31.2Claude Fable 5.1
- 31.2Muse Wonder
- 28.4GPT-6 Astra
- 27.6Grok 4.7
- 25.2GPT-6.1 Sol
- 23.6Gemini 4 Argon
- 22Claude Opus 5
- 22Muse Spark 1.3
15 modelsFull board
APEX-1 GP
- 73.04Claude Opus 5.5
- 72.12Claude Opus 5
- 68.36Gemini 3.8 Flash
- 64.35Kimi K3
- 63.75GPT-5.6 Sol
- 63.2GLM 5.3
- 61.03GPT-6 Astra
- 60.61GPT-6.1 Sol
- 59.38Grok 4.6
- 59.16Grok 4.7
12 modelsFull board
MLCR-AA
- 41.7Claude Sonnet 5.5
- 41.7GPT-6 Astra
- 38.3GPT-6.1 Sol
- 38.3Kimi K3.1 Pro
- 35.0Claude Haiku 5.5
- 33.9GPT-6 Sol
- 31.7GPT-5.5 Instant
- 30.6GPT-6.1 Terra
- 27.8Claude Opus 5.5
- 26.7DeepSeek V4.1 Flash 1017
27 modelsFull board
Bars are the model's lab colour. Each board's page says who ran every result.
Every model on every boardIndex against price
48 models on 3 or more boards with a published list price. Log price scale. The dashed line joins models that no cheaper model beats.
How the index is made
Each model's page shows its full calculation.
- One scale. Each result is read on its own board's 0 to 100 scale.
- Easy and hard boards. Each board's scores are read against how the same models do elsewhere, so a model is never marked down for a board it has no result on.
- Weights. Boards that measure clinical work most directly count most; boards near their ceiling count less.
- Coverage. A model with few boards is pulled toward the field average of 44.2, so the number of boards appears next to every score.
Each index score, with the board results beneath it
A forest plot of the Clinical Benchmarks Index. Each row is one model. Squares are its results on the index boards, each on that board's own 0 to 100 scale; the diamond is its index score. Ranks are across all 128 models.
- 1Claude Sonnet 5.5Anthropic · 7 of 12 boards63.27/12
- 2Claude Opus 5.5Anthropic · 8 of 12 boards62.88/12
- 3Claude Opus 5Anthropic · 10 of 12 boards56.810/12
- 4Claude Fable 5.1Anthropic · 6 of 12 boards56.86/12
- 5GPT-6 AstraOpenAI · 7 of 12 boards56.17/12
- 6Claude Haiku 5.5Anthropic · 4 of 12 boards55.54/12
- 7GPT-6.1 SolOpenAI · 7 of 12 boards53.97/12
- 8Muse Spark 1.1Meta · 3 of 12 boards52.73/12
- 9Kimi K3.1 ProMoonshot AI · 1 of 12 boards52.41/12
- 10Gemini 4 ArgonGoogle · 6 of 12 boards51.86/12
- 11GPT-5.5OpenAI · 5 of 12 boards51.75/12
- 12GPT-6.1 TerraOpenAI · 2 of 12 boards51.72/12
- 13Muse WonderMeta · 3 of 12 boards51.53/12
- 14GPT-5.6 SolOpenAI · 9 of 12 boards51.49/12
- 15Claude Fable 5Anthropic · 6 of 12 boards50.86/12
- 16GPT-6 LunaOpenAI · 6 of 12 boards50.66/12
- 17Claude Opus 4.8Anthropic · 4 of 12 boards50.64/12
- 18Kimi K3Moonshot AI · 7 of 12 boards50.47/12
- 19Gemini 3.7 FlashGoogle · 2 of 12 boards50.12/12
- 20Grok 4.7SpaceX AI · 7 of 12 boards50.07/12
- 21Claude Sonnet 5Anthropic · 6 of 12 boards49.86/12
- 22Gemini 3.5 FlashGoogle · 2 of 12 boards49.72/12
- 23GPT-6 SolOpenAI · 7 of 12 boards49.67/12
- 24Muse SparkMeta · 4 of 12 boards49.64/12
- 25Muse Wonder FullMeta · 2 of 12 boards49.52/12
- 26Gemini 3.6 FlashGoogle · 3 of 12 boards49.13/12
- 27GPT-5.6 TerraOpenAI · 4 of 12 boards48.94/12
- 28Gemini 3 Pro (11/25)Google · 1 of 12 boards48.71/12
- 29GLM 5.3 FlashZhipu · 1 of 12 boards48.61/12
- 30Gemini 3 FlashGoogle · 2 of 12 boards48.52/12
- 31DeepSeek V4.1 Flash 1017DeepSeek · 2 of 12 boards48.22/12
- 32Claude Opus 4.6Anthropic · 5 of 12 boards48.05/12
- 33MiniMax M3MiniMax · 2 of 12 boards48.02/12
- 34Claude Opus 4.5Anthropic · 3 of 12 boards47.73/12
- 35GLM 5.3Zhipu · 7 of 12 boards47.47/12
- 36Gemini 3.8 FlashGoogle · 7 of 12 boards47.37/12
- 37Claude Opus 4.7Anthropic · 4 of 12 boards47.34/12
- 38GPT-5.6 LunaOpenAI · 4 of 12 boards47.34/12
- 39Muse Spark 1.3Meta · 1 of 12 boards47.31/12
- 40Qwen3.8 MaxAlibaba · 4 of 12 boards47.04/12
- 41Grok 4.5SpaceX AI · 2 of 12 boards46.82/12
- 42Claude Sonnet 4.5Anthropic · 2 of 12 boards46.72/12
- 43Ember-1Fireworks AI · 1 of 12 boards46.51/12
- 44o3OpenAI · 2 of 12 boards46.42/12
- 45Gemini 2.5 FlashGoogle · 1 of 12 boards46.21/12
- 46Qwen3.8 27BAlibaba · 1 of 12 boards46.21/12
- 47GPT-5OpenAI · 4 of 12 boards46.14/12
- 48Qwen3.7 PlusAlibaba · 1 of 12 boards46.11/12
- 49GPT-5 miniOpenAI · 3 of 12 boards45.93/12
- 50Claude Opus 4.1Anthropic · 2 of 12 boards45.92/12
- 51Gemini 2.5 ProGoogle · 3 of 12 boards45.73/12
- 52Qwen3.5 397B A17BAlibaba · 3 of 12 boards45.73/12
- 53MiMo V2.6 FlashXiaomi · 2 of 12 boards45.62/12
- 54Grok 4.6SpaceX AI · 6 of 12 boards45.56/12
- 55GPT-5.2OpenAI · 4 of 12 boards45.54/12
- 56Gemini 3 ProGoogle · 2 of 12 boards45.52/12
- 57DeepSeek V4.1 FlashDeepSeek · 6 of 12 boards45.36/12
- 58GLM 5.2Zhipu · 2 of 12 boards45.22/12
- 59DeepSeek V4 FlashDeepSeek · 1 of 12 boards45.21/12
- 60Muse Spark 1.2Meta · 3 of 12 boards44.93/12
- 61GPT-5.4OpenAI · 5 of 12 boards44.55/12
- 62GPT-5.5 InstantOpenAI · 2 of 12 boards44.42/12
- 63DeepSeek V4 Flash 0731DeepSeek · 1 of 12 boards44.41/12
- 64Gemini 3.1 Flash Lite PreviewGoogle · 2 of 12 boards44.22/12
- 65Beam 1.0Reflection AI · 1 of 12 boards44.21/12
- 66MiMo V2.6 ProXiaomi · 3 of 12 boards44.13/12
- 67GPT-5.4 nanoOpenAI · 2 of 12 boards44.12/12
- 68Inkling SmallThinking Machines · 2 of 12 boards44.12/12
- 69Gemini 2.5 Flash (7/17)Google · 1 of 12 boards44.01/12
- 70Gemini 2.5 Flash Preview (9/25)Google · 1 of 12 boards44.01/12
- 71Gemini 3.5 Flash LiteGoogle · 2 of 12 boards43.92/12
- 72MiniMax M3.1 FlashMiniMax · 1 of 12 boards43.81/12
- 73Qwen 3.7 MaxAlibaba · 2 of 12 boards43.62/12
- 74Grok 4.3SpaceX AI · 3 of 12 boards43.53/12
- 75Grok 4 Fast (Reasoning)SpaceX AI · 2 of 12 boards43.52/12
- 76Kimi K2.5Moonshot AI · 3 of 12 boards43.33/12
- 77Nemotron 3 UltraNVIDIA · 1 of 12 boards43.31/12
- 78GPT-5.1OpenAI · 3 of 12 boards43.23/12
- 79Grok 4SpaceX AI · 2 of 12 boards43.22/12
- 80Ling 3.1 Flash 1019Ant Group · 1 of 12 boards43.11/12
- 81Gemini 3.1 ProGoogle · 6 of 12 boards42.46/12
- 82Nemotron 3.5 Super VLNVIDIA · 1 of 12 boards42.41/12
- 83Hy4 PreviewTencent · 3 of 12 boards42.33/12
- 84GLM 5.1Zhipu · 3 of 12 boards42.23/12
- 85Gemma 4 31BGoogle · 1 of 12 boards42.21/12
- 86MiniMax-M2.1MiniMax · 2 of 12 boards42.12/12
- 87Muse Spark OpenMeta · 2 of 12 boards42.02/12
- 88K2 Horizon 375B A23BInstitute of Foundation Models · 1 of 12 boards42.01/12
- 89DeepSeek V4 ProDeepSeek · 5 of 12 boards41.85/12
- 90DeepSeek R1DeepSeek · 1 of 12 boards41.61/12
- 91Claude Sonnet 4.6Anthropic · 3 of 12 boards41.53/12
- 92Gemini 2.5 Flash Lite (9/25)Google · 1 of 12 boards41.51/12
- 93Ling 3.0 FlashAnt Group · 2 of 12 boards41.42/12
- 94Gemma 4 26B A4BGoogle · 1 of 12 boards41.01/12
- 95Claude Sonnet 4Anthropic · 2 of 12 boards40.92/12
- 96InklingThinking Machines · 3 of 12 boards40.63/12
- 97Qwen 3 VL PlusAlibaba · 2 of 12 boards40.52/12
- 98Qwen 3.8 27BAlibaba · 2 of 12 boards40.52/12
- 99Grok 4 Fast (Non-Reasoning)SpaceX AI · 2 of 12 boards40.32/12
- 100Kimi K2.6Moonshot AI · 4 of 12 boards39.94/12
- 101o4-miniOpenAI · 2 of 12 boards39.82/12
- 102Qwen 3.5 FlashAlibaba · 2 of 12 boards39.82/12
- 103MiMo V2.5Xiaomi · 2 of 12 boards39.62/12
- 104Mistral Medium 3.5Mistral · 2 of 12 boards39.62/12
- 105Qwen 3 Max ThinkingAlibaba · 2 of 12 boards39.52/12
- 106GLM 4.7Zhipu · 2 of 12 boards39.32/12
- 107Grok 4.1 Fast (Reasoning)SpaceX AI · 2 of 12 boards39.32/12
- 108GPT-5 nanoOpenAI · 2 of 12 boards39.22/12
- 109Grok 4.1 Fast Non-ReasoningSpaceX AI · 2 of 12 boards39.22/12
- 110Ling 3.0 Flash FinAnt Group · 2 of 12 boards39.22/12
- 111MiMo V2.5 ProXiaomi · 3 of 12 boards38.63/12
- 112MAI-Thinking-1Microsoft · 1 of 12 boards38.51/12
- 113Gemini 2.5 Flash LiteGoogle · 2 of 12 boards38.42/12
- 114Llama 4 MaverickMeta · 2 of 12 boards38.12/12
- 115Qwen3.6 PlusAlibaba · 4 of 12 boards37.94/12
- 116Gemma 4 12BGoogle · 1 of 12 boards37.21/12
- 117Qwen3-VL-235B-A22BAlibaba · 1 of 12 boards36.81/12
- 118Mercury 2.5Inception · 2 of 12 boards36.32/12
- 119MiniMax-M2.7MiniMax · 3 of 12 boards35.33/12
- 120Laguna M.1Poolside · 2 of 12 boards35.12/12
- 121Laguna XS.2Poolside · 2 of 12 boards33.62/12
- 122Grok 4.20SpaceX AI · 4 of 12 boards32.64/12
- 123Llama 4 ScoutMeta · 2 of 12 boards32.42/12
- 124Command A+Cohere · 2 of 12 boards31.92/12
- 125Claude Haiku 4.5Anthropic · 5 of 12 boards31.05/12
- 126Gemma 4 E4BGoogle · 1 of 12 boards29.21/12
- 127Gemma 4 E2BGoogle · 1 of 12 boards27.11/12
- 128Nemotron 3.5 LightningNVIDIA · 1 of 12 boards14.71/12
Method of 31 October 2026. Select a square to read that result's quote and page.
How it is computed Index as CSVEvery board, on its own scale
The top 15 results on each board, in that board's rank order and on its own scale. The panels do not share an axis because the boards measure different things.
Bars are the model's lab colour. Each board's page says who ran every result.
All boards with detailsWhere the numbers come from
Each count is a count of rows in the published dataset.
source documents, each held with a verbatim quote and a page locator
results read from those documents
clinical boards tracked
- HealthAgentBench: fewer than 5 models
- CHI-Bench: fewer than 5 models
models ranked by the index
A result is added only with its source document, a verbatim quote and a locator. Results reported by the model's own lab are included and drawn hollow throughout.
Every source document