Clinical Benchmarks
Menu

Bedside Bench

Physician-written clinical cases on reasoning and safety, from differential diagnosis to medication safety.

Official pageReferences

More about this boardLess

Five hundred physician-authored clinical cases across ten suites of medical reasoning and safety, from differential diagnosis and workup to escalation and medication safety, each graded by an LLM judge against the authoring physicians' criteria.

Published by Doximity, released Oct 2026. 500 physician-written clinical cases in ten suites. Macro-average score % (avg@3), higher better.

Run by
task set held privately with Fireworks, which runs the evaluation
Headline metric
Score (avg@3)
Task count
500
Board last updated
22 Oct 2026
Publisher
Doximity
Rows
15
Models
15 of 16 on the board
Released
Oct 2026
Measured to
Oct 2026
Unit
500 clinical cases
Documents cited
1

Every result on Bedside Bench, on the board's own scale

15 results, one column each, highest first. Macro-average score % (avg@3), higher better.

15 of 15 results
  1. 94.6Claude Opus 5
  2. 94.3Muse Spark Open
  3. 93.4GPT-6 Astra
  4. 92.7Gemini 4 Argon
  5. 91.1GPT-5.6 Sol
  6. 89.7Kimi K3
  7. 89.1Ember-1
  8. 88.1GPT-6 Sol
  9. 87.2GPT-6 Luna
  10. 86.9GLM 5.3
  11. 86.5Gemini 3.8 Flash
  12. 85.65DeepSeek V4.1 Flash (1017)
  13. 83.3Grok 4.6
  14. 82.9DeepSeek V4.1 Flash
  15. 80.6DeepSeek V4 Pro 0813

Column colour is the model's lab. The table below says who ran each result.

Every published result, ranked on this board alone

15 rows. Higher is better. Bars run from 0 to 100; whiskers show a published interval. Select a row for its quote, page and document.

1Claude Opus 5Anthropic · Score94.6Board publisherOct 2026[1]
Claude Opus 5 94.6 (Score (avg@3)), rank 2
As printed
94.6
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
2Muse Spark OpenMeta · Score94.3Board publisherOct 2026[1]
Muse Spark Open 94.3 (Score (avg@3)), rank 3
As printed
94.3
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
3GPT-6 AstraOpenAI · Score93.4Board publisherOct 2026[1]
GPT-6 Astra 93.4 (Score (avg@3)), rank 4
As printed
93.4
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
4Gemini 4 ArgonGoogle · Score92.7Board publisherOct 2026[1]
Gemini 4 Argon 92.7 (Score (avg@3)), rank 5
As printed
92.7
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
5GPT-5.6 SolOpenAI · Score91.1Board publisherOct 2026[1]
GPT-5.6 Sol 91.1 (Score (avg@3)), rank 6
As printed
91.1
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
6Kimi K3Moonshot AI · Score89.7Board publisherOct 2026[1]
Kimi K3 89.7 (Score (avg@3)), rank 7
As printed
89.7
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
7Ember-1Fireworks AI · Score89.1Board publisherOct 2026[1]
Ember-1 89.1 (Score (avg@3)), rank 8
As printed
89.1
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
8GPT-6 SolOpenAI · Score88.1Board publisherOct 2026[1]
GPT-6 Sol 88.1 (Score (avg@3)), rank 9
As printed
88.1
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
9GPT-6 LunaOpenAI · Score87.2Board publisherOct 2026[1]
GPT-6 Luna 87.2 (Score (avg@3)), rank 10
As printed
87.2
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
10GLM 5.3Zhipu · Score86.9Board publisherOct 2026[1]
GLM 5.3 86.9 (Score (avg@3)), rank 11
As printed
86.9
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
11Gemini 3.8 FlashGoogle · Score86.5Board publisherOct 2026[1]
Gemini 3.8 Flash 86.5 (Score (avg@3)), rank 12
As printed
86.5
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
12DeepSeek V4.1 Flash (1017)DeepSeek · Score85.65Board publisherOct 2026[1]
DeepSeek V4.1 Flash (1017) 85.65 (Score (avg@3)), rank 13
As printed
85.65
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
13Grok 4.6SpaceX AI · Score83.3Board publisherOct 2026[1]
Grok 4.6 83.3 (Score (avg@3)), rank 14
As printed
83.3
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
14DeepSeek V4.1 FlashDeepSeek · Score82.9Board publisherOct 2026[1]
DeepSeek V4.1 Flash 82.9 (Score (avg@3)), rank 15
As printed
82.9
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
15DeepSeek V4 Pro 0813DeepSeek · Score80.6Board publisherOct 2026[1]
DeepSeek V4 Pro 0813 80.6 (Score (avg@3)), rank 16
As printed
80.6
Where
https://fireworks.ai/specialized-intelligence-index/benchmarks/healthcare/bedside-bench (feed: benchmarks[id=healthcare/bedside-bench].results)
Setting
Score (avg@3)
Reported by
official leaderboard
Document
[1] Bedside Bench on the Fireworks Specialized Intelligence Index
References

1 document supports this board

  1. 1.Bedside Bench on the Fireworks Specialized Intelligence Indexofficial leaderboard, Fireworks AI, 22 Oct 2026. Every result it supports