Benchmarks / Knowledge

MedCode

Can models support the medical billing process?

95models
–Open Index weight
2026-10-13last updated
Vals AIpublisher

MedCode

Higher is better · best effort per model

Chart settings
95 of 95 selected

full table below

#ModelEffortScoreCost / taskHarness
1Opus 5max63.6%$0.157–
2Gemini 3.1 Prohigh59.1%$0.025–
3Gemini 4high58.8%$0.137–
4Fable 5max56.1%$0.591–
5Gemini 3 Flashhigh55.9%$0.0062–
6Gemini 3.5 Flashhigh55.8%$0.074–
7Opus 4.7max54.9%$0.226–
8Fable 5.1max53.5%$1.12–
9Gemini 3.7 Flashhigh53.4%$0.038–
10Opus 4.8max53.2%$0.351–
11Gemini 3.6 Flashhigh53.2%$0.044–
12Sonnet 5.5max52.9%$0.392–
13GPT-5.1high52.7%$0.014–
14Gemini 3 Prohigh52.2%$0.028–
15Haiku 5.5max51.5%$0.032–
16Muse Spark–51.3%$0.0053–
17Gemini 2.5 Prohigh50.6%$0.015–
18Wonder Fullmax49.9%$0.247–
19Opus 5.5max49.8%$0.658–
20GPT-5.2xhigh49.7%$0.019–
21GPT-5high49.6%$0.046–
22Grok 4.7xhigh49.6%$0.105–
23Kimi K3max49.4%$0.094–
24Muse Spark 1.2xhigh49.3%$0.039–
25Opus 4.5high49.2%$0.096–
26Opus 4.6max49.1%$0.244–
27GPT-5.5xhigh49.1%$0.161–
28GPT-6.1 Solmax48.8%$0.085–
29GPT-6 Astramax48.5%$0.451–
30Gemini 3.8 Flashhigh48.1%$0.018–
31Gemini 3.1 Flash-Litehigh47.6%$0.0020–
32Sonnet 5max47.5%$0.279–
33o3high47.3%$0.030–
34Opus 4.1high47.2%$0.269–
35GPT-6 Solmax47.1%$0.086–
36MiniMax M3max46.3%$0.012–
37MiMo V2.6 Prohigh45.0%$0.0095–
38Grok 4.6high44.7%$0.050–
39GPT-6 Lunamax44.7%$0.0073–
40Sonnet 4.5high44.1%$0.101–
41GPT-5.6 Solmax44.0%$0.281–
42Gemini 3.5 Flash-Litehigh43.5%$0.0081–
43GPT-5.6 Terraxhigh43.4%$0.047–
44Grok 4.5high43.3%$0.048–
45Hy4–43.2%$0.062–
46GPT-5 Minihigh43.0%$0.0056–
47GLM 5.3max42.9%$0.082–
48DeepSeek V4 Pro 0813max42.5%$0.061–
49GPT-5.6 Lunamax42.4%$0.016–
50GLM 5.1max41.6%$0.024–
51DeepSeek V4 Flash 0731high41.4%$0.020–
52GPT-5.4xhigh41.3%$0.212–
53Inklingmax41.2%$0.127–
54DeepSeek V4.1 Flashhigh41.2%$0.012–
55MiMo V2.6 Flashhigh41.1%$0.0029–
56GPT-5.4 Nanohigh41.0%$0.0008–
57GLM 5.2max40.8%$0.045–
58Mistral Large 4high40.7%$0.042–
59Qwen 3.8 Maxmax40.7%$0.121–
60Gemini 2.5 Flashmedium40.5%$0.0037–
61DeepSeek V4max40.5%$0.061–
62Kimi K2.6max40.1%$0.041–
63Kimi K2.5max39.3%$0.017–
64Qwen 3.7 Maxmax38.8%$0.042–
65Nemotron 3 Ultra 550B A55Bmax38.6%––
66Grok 4high38.1%$0.034–
67Grok 4.3xhigh38.1%$0.022–
68Inkling Smallxhigh37.9%$0.017–
69Grok 4 Fastnone37.4%$0.0021–
70Qwen 3.6 Plusmax36.9%$0.016–
71Llama 4 Maverickmedium36.5%$0.0029–
72Sonnet 4high35.0%$0.070–
73MiniMax M2.7high34.4%$0.0074–
74Gemini 2.5 Flash Litenone34.2%$0.0012–
75MiniMax M2.1–34.1%––
76o4 minihigh33.8%$0.018–
77Mistral Medium 3.5high33.8%$0.053–
78Qwen 3.5 Flashmax33.0%$0.0039–
79GLM 4.7none32.8%$0.0067–
80Haiku 4.5xhigh32.7%$0.020–
81MiMo V2.5 Proxhigh32.5%$0.0067–
82Ling 3.0 Flash–32.3%$0.0016–
83Grok 4.20xhigh32.2%$0.036–
84MiMo V2.5xhigh31.9%$0.0022–
85Qwen 3high31.7%$0.0025–
86Qwen 3 Maxmax31.4%$0.015–
87Mercury 2.5high31.3%$0.0045–
88GPT-5 Nanohigh30.4%$0.0017–
89Ling 3.0 Flash Fin–29.3%$0.0014–
90Qwen 3.8 27Bxhigh28.7%$0.050–
91Grok 4.1 Fasthigh28.3%$0.0022–
92Llama 4 Scoutmedium23.3%$0.0022–
93Laguna M.1–23.1%$0.0030–
94Laguna XS.2–21.3%$0.0014–
95Command Anone19.7%$0.056–