Vals MedCode (MedCode)

We show this table for reference; we do not rank on it.

Data verified September 26, 202636 confirmed releases in the last 30 daysFollow model changes

Vals AI healthcare benchmark for whether models can support the medical billing process.

MedCode score on MedCode — September 26, 2026

We mirror the published medcode score view for MedCode. Claude Opus 5 leads the public snapshot at 63.57%, followed by Gemini 3.1 Pro Preview (59.06%) and Claude Fable 5 (56.07%). We do not use these results to rank models overall.

1Proprietary

Claude Opus 5

Anthropic

63.57%

Overall 79.24

Gemini 3.1 Pro Preview

Google

59.06%

3Proprietary

Claude Fable 5

Anthropic

56.07%

Overall 78.59Context 1M+

102 modelsAgenticCurrentDisplay onlyUpdated September 26, 2026

MedCode score table (102 models)

Score

Claude Opus 5Anthropic · Closed

63.57%

Gemini 3.1 Pro PreviewGoogle

59.06%

Claude Fable 5Anthropic · Closed

56.07%

Gemini 3 Flash PreviewGoogle

55.92%

Gemini 3.5 FlashGoogle · Closed

55.83%

Claude Opus 4.7Anthropic · Closed

54.86%

Claude Fable 5.1Anthropic · Closed

53.51%

Gemini 3.7 FlashGoogle · Closed

53.39%

Claude Opus 4.8Anthropic · Closed

53.22%

Gemini 3.6 FlashGoogle · Closed

53.15%

Claude Sonnet 5.5Anthropic · Closed

52.92%

GPT-5.1OpenAI · Closed

52.73%

Gemini 3 Pro PreviewGoogle

52.20%

Muse SparkMeta · Closed

51.31%

Gemini 2.5 ProGoogle · Closed

50.59%

Claude Opus 5.5Anthropic · Closed

49.80%

GPT-5.2OpenAI · Closed

49.75%

GPT-5OpenAI

49.63%

Grok 4.7xAI · Closed

49.55%

Muse Spark 1.2Meta · Closed

49.35%

Claude Opus 4.5 ThinkingAnthropic · Closed

49.16%

Claude Opus 4.6 (Adaptive)Anthropic · Closed

49.13%

GPT-5.5OpenAI · Closed

49.10%

Kimi K3Moonshot AI · Closed

48.88%

GPT-6 AstraOpenAI · Closed

48.49%

Claude Opus 4.6Anthropic · Closed

48.24%

Gemini 3.8 FlashGoogle · Closed

48.13%

Gemini 3.1 Flash Lite PreviewGoogle

47.60%

Claude Sonnet 5Anthropic · Closed

47.54%

o3OpenAI · Closed

47.29%

Claude Opus 4.1 20250805 ThinkingAnthropic

47.23%

GPT-6 SolOpenAI · Closed

47.07%

MiniMax M3MiniMax · Open weight

46.29%

Claude Opus 4.5Anthropic · Closed

45.17%

MiMo-V2.6-ProXiaomi · Open weight

44.97%

Grok 4.6xAI · Closed

44.71%

GPT-6 LunaOpenAI · Closed

44.69%

Claude Sonnet 4.5 ThinkingAnthropic · Closed

44.13%

GPT-5.6 SolOpenAI · Closed

43.97%

Gemini 3.5 Flash-LiteGoogle · Closed

43.49%

GPT-5.6 TerraOpenAI · Closed

43.41%

Grok 4.5xAI · Closed

43.29%

Hy4 previewTencent · Open weight

43.25%

GPT-5 miniOpenAI · Closed

43.05%

GLM-5.3Z.AI · Open weight

42.86%

DeepSeek V4 Pro 0813DeepSeek · Open weight

42.47%

GPT-5.6 LunaOpenAI · Closed

42.39%

GLM-5.1Z.AI · Open weight

41.60%

DeepSeek V4 Flash 0731DeepSeek · Open weight

41.41%

Claude Opus 4.1Anthropic

41.37%

GPT-5.4OpenAI · Closed

41.29%

InklingThinking Machines Lab · Open weight

41.19%

DeepSeek V4.1 FlashDeepSeek · Open weight

41.17%

MiMo-V2.6-FlashXiaomi · Open weight

41.06%

GPT-5.4 nanoOpenAI · Closed

41.03%

GLM-5.2Z.AI · Open weight

40.77%

Qwen3.8 MaxAlibaba · Open weight

40.67%

Claude Sonnet 4.5Anthropic · Closed

40.57%

Gemini 2.5 Flash Preview 09 2025Google

40.54%

DeepSeek V4 Pro 0813DeepSeek · Open weight

40.45%

Gemini 2.5 Flash ThinkingGoogle

40.36%

Gemini 2.5 Flash Preview 09 2025 ThinkingGoogle

40.33%

Kimi K2.6Moonshot AI · Open weight

40.14%

Kimi K2.5 ThinkingMoonshot AI

39.32%

Qwen3.7 MaxAlibaba · Closed

38.75%

Nemotron 3 Ultra 550b A55bnvidia

38.62%

Gemini 2.5 FlashGoogle · Closed

38.42%

Grok 4 0709xAI

38.08%

Grok 4.3xAI · Closed

38.07%

Inkling-SmallThinking Machines Lab · Open weight

37.89%

Grok 4 Fast (Reasoning)xAI · Closed

37.38%

Qwen3.6 PlusAlibaba · Closed

36.89%

Llama4 Maverick Instruct BasicFireworks

36.51%

Claude Sonnet 4 20250514 ThinkingAnthropic

34.96%

MiniMax M2.7MiniMax · Open weight

34.44%

Gemini 2.5 Flash Lite Preview 09 2025 ThinkingGoogle

34.19%

MiniMax M2.1MiniMax

34.08%

Claude Sonnet 4Anthropic

33.94%

O4 MiniOpenAI

33.79%

Mistral Medium 3.5Mistral

33.75%

Qwen3.5 FlashAlibaba · Closed

33.00%

GLM-4.7Z.AI · Open weight

32.77%

Claude Haiku 4.5 ThinkingAnthropic · Closed

32.68%

MiMo-V2.5-ProXiaomi · Closed

32.48%

Ling 3.0 Flash 2607ant

32.27%

Grok 4.20 0309 ReasoningxAI

32.16%

MiMo-V2.5Xiaomi · Closed

31.89%

Qwen3 VL PlusAlibaba

31.65%

Qwen3 MaxAlibaba · Closed

31.37%

Mercury 2.5Inception · Closed

31.33%

GPT-5 nanoOpenAI · Closed

30.44%

Grok 4 Fast Non ReasoningxAI

30.04%

Ling 3.0 Flash Af Rc3ant

29.30%

Qwen3.8-27BAlibaba · Open weight

28.70%

Grok 4.1 Fast Non ReasoningxAI

28.35%

Grok 4.1 Fast (Reasoning)xAI · Closed

28.08%

Gemini 2.5 Flash LiteGoogle

27.11%

Gemini 2.5 Flash Lite Preview 09 2025Google

27.08%

Meta Llama Llama 4 Scout 17B 16E InstructTogether AI

23.31%

100

Laguna M.1Poolside · Closed

23.11%

101

Laguna XS.2Poolside · Open weight

21.25%

102

Command A Plus 05 2026Cohere

19.72%

How MedCode is shown here

BenchLM mirrors the public Vals AI MedCode leaderboard captured from https://www.vals.ai/benchmarks/medcode and updated by Vals on September 26, 2026. The snapshot preserves overall scores, uncertainty, latency, cost-per-test metadata, and task-level scores where Vals publishes them.

MedCode is display only on BenchLM. Vals proprietary or Vals-hosted aggregate views are useful context, but BenchLM does not use them as weighted ranking inputs or as a replacement for benchmark-native source records.

Snapshot

102 Vals rows1 task viewsprivate datasetTasks: OverallDisplay only

MedCode on Vals AI Vals methodology Vals home

The published MedCode snapshot places Claude Opus 5 first at 63.57%. The third row is 7.50 points behind. The broader top-10 range is 10.42 points, so the table still separates the published systems.

102 models have been evaluated on MedCode. The benchmark falls in the Agentic category. We keep external benchmark mirrors separate from the weighted global scoring system, so these results remain source-specific evidence. MedCode is currently displayed for reference but excluded from the scoring formula, so it does not directly affect overall rankings.

About MedCode

Year

2026

Tasks

Medical billing support tasks

Format

Accuracy score

Difficulty

Professional healthcare administration

BenchLM mirrors the public Vals MedCode leaderboard as display-only healthcare evidence.

MedCode Public benchmark source

Freshness and provenance

Version

MedCode 2026

Refresh cadence

Quarterly

Staleness state

Current

Question availability

Public benchmark set

CurrentDisplay only

BenchLM uses freshness metadata to decide whether a benchmark should still be treated as a strong differentiator, a benchmark to watch, or a display-only reference. For the full scoring policy, see the BenchLM methodology page.

Questions

What does MedCode measure?

Vals AI healthcare benchmark for whether models can support the medical billing process.

Which model leads the published MedCode snapshot?

Claude Opus 5 currently leads the published MedCode snapshot with 63.57% medcode score. BenchLM shows this benchmark for display only and does not use it in overall rankings.

How many models are evaluated on MedCode?

The September 26, 2026 snapshot contains 102 AI models.

Last updated: September 26, 2026 · mirrored from the public benchmark leaderboard

Know when it’s worth switching models

The model to choose, the cheaper alternative, and the release we would wait on.

Read a sample issue

Join 2,000+ readers.

One email each week. Unsubscribe anytime.