BenchLeader
Menu
Best LLMsChoose a modelCoding plansLeaderboardsModelsBenchmarksProvidersReleases7

MedScribe leaderboard

As of 6 Nov 2026, the top score on MedScribe is 91.4% by Claude Opus 5.5 (Anthropic), ahead of Claude Fable 5.1 at 91.3%, across 107 model configurations with a published result (Vals AI, result dated 29 Oct 2026). From an encounter transcript the model drafts the paperwork a doctor's office produces, graded against reference documents. MedScribe is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 6 Nov 2026)

  • Claude Opus 5.5 holds the top MedScribe score at 91.4% (Vals AI, 29 Oct 2026).
  • Claude Fable 5.1 is second on MedScribe at 91.3% (Vals AI, 29 Oct 2026).
  • MedScribe: Doctor administrative paperwork from encounter notes. Run by Vals AI (Vals AI).
  • 107 model configurations have a published MedScribe result on BenchLeader (BenchLeader, 6 Nov 2026).
  • MedScribe does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Doctor administrative paperwork from encounter notes. Run by Vals AI.

Published by
Vals AI
Category
Knowledge
Index weight
Reference only
Models
107
Data as of
6 Nov 2026

Vals AI (vals.ai).

What the test looks like

From an encounter transcript the model drafts the paperwork a doctor's office produces, graded against reference documents.

How it is scored

Rubric score, run by Vals AI.

What to keep in mind

Rubric-graded writing; style choices affect scores.

Top 15
  1. 1Claude Opus 5.591.4%
  2. 2Claude Fable 5.191.3%
  3. 3Claude Sonnet 5.591.1%
  4. 4Claude Opus 591.0%
  5. 5Muse Spark 1.2 (xhigh)90.1%
  6. 6Claude Haiku 5.589.5%
  7. 7Grok 4.7 (xhigh)89.4%
  8. 8Muse Wonder Full (max)89.3%
  9. 9GLM 5.3 Flash (max)88.9%
  10. 10Muse Spark 1.1 (xhigh)88.9%
  11. 11GLM 5.3 (max)88.8%
  12. 12Claude Fable 588.5%
  13. 13MiMo-V2.6-Pro88.3%
  14. 14GPT-5.1 (high)88.1%
  15. 15Kimi K3 (max)88.0%
107 of 107
#
1Claude Opus 5.591.4%63.4$1.152026-10-29
2Claude Fable 5.191.3%66.4$0.9642026-10-29
3Claude Sonnet 5.591.1%–$0.5092026-10-29
4Claude Opus 591.0%65.2$0.2362026-10-29
5Muse Spark 1.2xhigh90.1%61.7$0.0382026-10-29
6Claude Haiku 5.589.5%–$0.0442026-10-29
7Grok 4.7xhigh89.4%58.3$0.0752026-10-29
8Muse Wonder Fullmax89.3%62.9$0.1922026-10-29
9GLM 5.3 Flashmax88.9%54.2$0.00222026-10-29
10Muse Spark 1.1xhigh88.9%59.5$0.0352026-10-29
11GLM 5.3max88.8%59.0$0.0572026-10-29
12Claude Fable 588.5%66.5$0.5832026-10-29
13MiMo-V2.6-Pro88.3%57.3$0.00992026-10-29
14GPT-5.1high88.1%57.7$0.0972026-10-29
15Kimi K3max88.0%62.1$0.1102026-10-29
16GPT-6 Astramax87.9%67.8$0.5822026-10-29
17Gemini 4 Argonhigh87.4%66.0$0.1472026-10-29
18MiniMax M387.3%56.4$0.0142026-10-29
19Grok 4.5high86.9%56.9$0.0332026-10-29
20GPT-5.5xhigh86.9%64.3$0.1432026-10-29
21Claude Opus 4.686.7%62.3$0.1152026-10-29
22Grok 4.6high86.5%60.5$0.0372026-10-29
23GPT-6.1 Solmax86.5%67.3$0.1012026-10-29
24Claude Opus 4.6thinking86.1%62.4$0.2252026-10-29
25Muse Spark85.9%64.1$0.00772026-10-29
26Claude Opus 4.885.8%60.0$0.2592026-10-29
27DeepSeek V4.1 Flashhigh85.5%55.7$0.0152026-10-29
28Inkling85.4%47.5$0.1662026-10-29
29Claude Opus 4.5thinking85.3%59.6$0.4102026-10-29
30MiMo-V2.6-Flash85.3%55.0$0.00222026-10-29
31GPT-5.6 Solmax85.2%65.4$0.2772026-10-29
32Claude Haiku 4.5thinking85.2%45.4$0.0422026-10-29
33Qwen3.8 Maxmax85.0%58.8$0.0902026-10-29
34Claude Sonnet 4.584.5%53.7$0.0552026-10-29
35Gemini 3.8 Flashhigh84.5%62.0$0.0252026-10-29
36GPT-5.2xhigh84.4%60.0$0.1152026-10-29
37GPT-5.6 Lunamax84.4%56.3$0.0232026-10-29
38Inkling Small84.1%52.1$0.0192026-10-29
39Claude Sonnet 4.5thinking84.1%52.2$0.0822026-10-29
40Gemini 3.7 Flashhigh83.9%62.2$0.0592026-10-29
41Qwen3.8 27Bxhigh83.8%52.5$0.0472026-10-29
42MiMo-V2.5-Pro83.7%54.9$0.00602026-10-29
43GPT-6 Lunamax83.7%56.7$0.00962026-10-29
44GPT-5high83.7%57.9$0.1022026-10-29
45Hy483.6%55.7$0.0532026-10-29
46GLM 5.283.5%51.5$0.0452026-10-29
47Claude Opus 4.583.3%57.9$0.2822026-10-29
48Gemini 2.5 Flashthinking83.0%48.3$0.0152026-10-29
49Claude Opus 4.783.0%63.1$0.1782026-10-29
50GPT-5.6 Terraxhigh82.9%60.3$0.0632026-10-29
51Gemini 2.5 Flash82.9%52.7$0.0152026-10-29
52GPT-6 Solmax82.0%63.3$0.0832026-10-29
53Grok 4 Fastthinking81.6%51.6$0.00252026-10-29
54Ling 3.0 Flash80.9%51.8$0.00142026-10-29
55MiniMax-M2.180.8%55.0$0.00512026-10-29
56GPT-5 minihigh80.6%54.6$0.0332026-10-29
57Mistral Large 4high80.4%49.5$0.0242026-10-29
58DeepSeek V4 Flashhigh80.4%57.3$0.0142026-10-29
59DeepSeek V4 Promax80.2%57.0$0.0412026-10-29
60MiniMax M2.779.9%56.5$0.00512026-10-29

More knowledge benchmarks · All LLM benchmarks

Related benchmarks

Top head-to-heads on MedScribe

Cite as: BenchLeader, “MedScribe leaderboard”, data as of 6 Nov 2026, https://www.benchleader.com/benchmarks/vals_medscribe.

MedScribe: questions

What does MedScribe measure?
From an encounter transcript the model drafts the paperwork a doctor's office produces, graded against reference documents. Scores are reported in percent of tasks solved; higher is better.
What is the top score on MedScribe?
As of 6 Nov 2026, the highest score on MedScribe is 91.4% by Claude Opus 5.5 (result dated 29 Oct 2026), ahead of Claude Fable 5.1 at 91.3%. Scores come from Vals AI and are re-read daily.
Which AI model leads MedScribe?
Claude Opus 5.5 leads MedScribe with 91.4% as of 6 Nov 2026 (result dated 29 Oct 2026), ahead of Claude Fable 5.1 at 91.3%.
How many models have MedScribe results?
107 model configurations have a MedScribe result on BenchLeader, all taken from Vals AI.
Who runs MedScribe and how often is it updated?
MedScribe is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
Does MedScribe count toward the BenchLeader Index?
No. MedScribe is shown for reference but left out of the composite index.