BenchLeader

MedCode leaderboard

As of 15 Nov 2026, the top score on MedCode is 63.6% by Claude Opus 5 (Anthropic), ahead of Gemini 3.1 Pro at 59.1%, across 104 model configurations with a published result (Vals AI, result dated 29 Oct 2026). Clinical notes are given and the model must assign the correct medical billing codes, a task with strict, checkable answers. MedCode is shown for reference only and does not count toward the BenchLeader Index.

Key facts (as of 15 Nov 2026)

  • Claude Opus 5 holds the top MedCode score at 63.6% (Vals AI, 29 Oct 2026).
  • Gemini 3.1 Pro is second on MedCode at 59.1% (Vals AI, 29 Oct 2026).
  • MedCode: Assigning medical billing codes from clinical notes. Run by Vals AI (Vals AI).
  • 104 model configurations have a published MedCode result on BenchLeader (BenchLeader, 15 Nov 2026).
  • MedCode does not feed the BenchLeader Index; it is shown for reference (BenchLeader methodology).

Assigning medical billing codes from clinical notes. Run by Vals AI.

Published by
Vals AI
Category
Knowledge
Index weight
Reference only
Models
104
Data as of
15 Nov 2026

Vals AI (vals.ai).

What the test looks like

Clinical notes are given and the model must assign the correct medical billing codes, a task with strict, checkable answers.

How it is scored

Accuracy, run by Vals AI.

What to keep in mind

Narrow but economically real; US coding systems only.

104 of 104
#
1Claude Opus 5max63.6%65.0$0.1572026-10-29
2Gemini 3.1 Prohigh59.1%60.4$0.0252026-10-29
3Gemini 4 Argonhigh58.8%65.3$0.1372026-10-29
4Claude Fable 5max56.1%65.1$0.5912026-10-29
5Gemini 3 Flashhigh55.9%57.4$0.00622026-10-29
6Gemini 3.5 Flashhigh55.8%62.5$0.0742026-10-29
7Claude Opus 4.7max54.9%59.7$0.2262026-10-29
8Claude Fable 5.1max53.5%66.7$1.122026-10-29
9Gemini 3.7 Flashhigh53.4%62.1$0.0382026-10-29
10Claude Opus 4.8max53.2%61.1$0.3512026-10-29
11Gemini 3.6 Flashhigh53.1%61.0$0.0442026-10-29
12Claude Sonnet 5.5max52.9%64.2$0.3922026-10-29
13GPT-5.1high52.7%57.5$0.0142026-10-29
14Gemini 3 Prohigh52.2%60.0$0.0282026-10-29
15Claude Haiku 5.5max51.5%55.3$0.0322026-10-29
16Muse Spark51.3%64.2$0.00532026-10-29
17Gemini 2.5 Pro50.6%55.2$0.0152026-10-29
18Muse Wonder Fullmax49.9%61.7$0.2472026-10-29
19Claude Opus 5.5max49.8%68.4$0.6342026-10-29
20GPT-5.2xhigh49.8%59.9$0.0192026-10-29
21GPT-5high49.6%58.2$0.0462026-10-29
22Grok 4.7xhigh49.5%57.4$0.1052026-10-29
23Kimi K3max49.4%60.2$0.0942026-10-29
24Muse Spark 1.2xhigh49.4%61.5$0.0392026-10-29
25Claude Opus 4.5high49.2%59.7$0.0962026-10-29
26Claude Opus 4.6max49.1%57.0$0.2442026-10-29
27GPT-5.5xhigh49.1%64.1$0.1612026-10-29
28GPT-6.1 Solmax48.8%66.6$0.0852026-10-29
29GPT-6 Astramax48.5%67.3$0.4512026-10-29
30Gemini 3.8 Flashhigh48.1%59.4$0.0182026-10-29
31Gemini 3.1 Flash Litehigh47.6%48.4$0.00202026-10-29
32Claude Sonnet 5max47.5%56.1$0.2792026-10-29
33o3high47.3%55.1$0.0302026-10-29
34Claude Opus 4.1thinking47.2%58.1$0.2692026-10-29
35GPT-6 Solmax47.1%62.6$0.0862026-10-29
36MiniMax M346.3%56.1$0.0122026-10-29
37MiMo-V2.6-Pro45.0%57.5$0.00952026-10-29
38Grok 4.6high44.7%60.6$0.0502026-10-29
39GPT-6 Lunamax44.7%55.1$0.00732026-10-29
40Claude Sonnet 4.5thinking44.1%52.3$0.1012026-10-29
41GPT-5.6 Solmax44.0%65.3$0.2812026-10-29
42Gemini 3.5 Flash Litehigh43.5%47.8$0.00812026-10-29
43GPT-5.6 Terraxhigh43.4%60.1$0.0472026-10-29
44Grok 4.5high43.3%56.8$0.0482026-10-29
45Hy443.3%56.7$0.0622026-10-29
46GPT-5 minihigh43.0%54.6$0.00562026-10-29
47GLM 5.3max42.9%58.8$0.0822026-10-29
48DeepSeek V4 Promax42.5%54.7$0.0612026-10-29
49GPT-5.6 Lunamax42.4%56.2$0.0162026-10-29
50GLM 5.141.6%55.4$0.0242026-10-29
51DeepSeek V4 Flashhigh41.4%57.2$0.0202026-10-29
52Claude Opus 4.141.4%51.1$0.2062026-10-29
53GPT-5.4xhigh41.3%63.1$0.2122026-10-29
54Inkling41.2%50.4$0.1272026-10-29
55DeepSeek V4.1 Flashhigh41.2%55.6$0.0122026-10-29
56MiMo-V2.6-Flash41.1%54.8$0.00292026-10-29
57GPT-5.4 Nanohigh41.0%48.3$0.00082026-10-29
58GLM 5.240.8%51.9$0.0452026-10-29
59Mistral Large 4high40.7%48.3$0.0212026-10-29
60Qwen3.8 Maxmax40.7%58.7$0.1212026-10-29

More knowledge benchmarks · All LLM benchmarks

Cite as: BenchLeader, “MedCode leaderboard”, data as of 15 Nov 2026, https://www.benchleader.com/benchmarks/vals_medcode.

MedCode: questions

What does MedCode measure?
Clinical notes are given and the model must assign the correct medical billing codes, a task with strict, checkable answers. Scores are reported in percent of tasks solved; higher is better.
What is the top score on MedCode?
As of 15 Nov 2026, the highest score on MedCode is 63.6% by Claude Opus 5 (result dated 29 Oct 2026), ahead of Gemini 3.1 Pro at 59.1%. Scores come from Vals AI and are re-read daily.
Which AI model leads MedCode?
Claude Opus 5 leads MedCode with 63.6% as of 15 Nov 2026 (result dated 29 Oct 2026), ahead of Gemini 3.1 Pro at 59.1%.
How many models have MedCode results?
104 model configurations have a MedCode result on BenchLeader, all taken from Vals AI.
Who runs MedCode and how often is it updated?
MedCode is published by Vals AI. BenchLeader re-reads the published results once a day and records the date each result was published.
Does MedCode count toward the BenchLeader Index?
No. MedCode is shown for reference but left out of the composite index.