MedCode
Professional Work · 2026-02-19
Healthcare benchmark testing whether models can support medical billing and coding workflows. Scores are reported as task accuracy.
Top models (higher is better)
| Model | Score |
|---|---|
| Claude Opus 5 | 63.6 |
| Gemini 3.1 Pro Preview | 59.1 |
| Claude Fable 5 | 56.1 |
| Gemini 3 Flash Preview | 55.9 |
| Gemini 3.5 Flash | 55.8 |
| Opus 4.7 | 54.9 |
| Opus 4.8 | 53.2 |
| Gemini 3.6 Flash | 53.2 |
| GPT-5.1 | 52.7 |
| Gemini 3 Pro Preview | 52.2 |
| Muse Spark | 51.3 |
| Gemini 2.5 Pro | 50.6 |
| GPT-5.2 | 49.7 |
| GPT-5 | 49.6 |
| Opus 4.5 | 49.2 |
| Opus 4.6 | 49.1 |
| GPT-5.5 | 49.1 |
| Kimi K3 | 48.9 |
| Gemini 3.1 Flash-Lite Preview | 47.6 |
| o3 | 47.3 |
| Opus 4.1 | 47.2 |
| MiniMax M3 | 46.3 |
| Sonnet 4.5 | 44.1 |
| GPT-5.6 Sol | 44.0 |
| Gemini 3.5 Flash-Lite | 43.5 |
| GPT-5.6 Terra | 43.4 |
| Grok 4.5 | 43.3 |
| GPT-5 Mini | 43.0 |
| GPT-5.6 Luna | 42.4 |
| GLM-5.1 | 41.6 |
| GPT-5.4 | 41.3 |
| Inkling | 41.2 |
| GPT-5.4 Nano | 41.0 |
| GLM-5.2 | 40.8 |
| Gemini 2.5 Flash Preview (09-2025) | 40.5 |
| DeepSeek-V4-Pro | 40.5 |
| Gemini 2.5 Flash | 40.4 |
| Kimi K2.6 | 40.1 |
| Kimi K2.5 | 39.3 |
| Qwen3.7-Max | 38.8 |