Commure RCM-Bench-1: leaderboard

Healthcare revenue-cycle management agent benchmark from Commure Engineering evaluating models across 97 clinical billing cases (3 trials per case, 291 total trials) spanning six workflows: claim submission, encounter billing, patient billing, insurance updates, eligibility checks, and RCM knowledge. Scored by equally weighted average pass rate across the six workflows with failed attempts retained in the denominator.

Metric: Weighted workflow pass rate (%, 6 RCM workflows, 97 cases x 3 trials). Source: eng.commure.com. Saturation forecast: Not forecast. 27 models tracked. Updated 20 Nov 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Fable 5.194.15#20
2Gemini 4 HydrogenMedium92.91#16
3Claude Opus 592.68#26
4Gemini 4 ArgonMedium91.49#18
5Claude Sonnet 591.24#118
6GPT-5.6 SolMedium91.21#36
7Claude Sonnet 5.591.05#24
8Gemini 4 CarbonMedium90.38#5
9Gemini 3.8 FlashMedium90.23#43
10Claude Opus 5.590.14#9
11Claude Sonnet 4.689.82#175
12DeepSeek V4.1 Flash88.99#69
13Claude Haiku 5.588.38#75
14GLM-5.388.3#73
15Grok 4.788.16#51

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon.

Interactive version: theaggregate.ai/benchmark?slug=commure-rcm-bench-1 · How It Works · Data refreshed daily, snapshot 2026-11-20.