HealthAdminBench

Healthcare administration agent benchmark for prior authorization, appeals, durable medical equipment, payer portals, fax, and EHR-adjacent workflows.

7rows
task_success_rateprimary metric
2026-05-27sampled

Metadata

ID
healthadminbench
Category
Healthcare
Release
Unknown

Metrics

Task Success Rate, Score, Max Score, Avg. Steps (lower is better), Avg. Time (lower is better)

Latest Results

Rows are aggregated from task-level public results for the website's default Task Description + Portal Guidance prompt and screenshot observation mode.

Rank Subject Task Success Rate Model Match Provenance Sampled
1 openai-cua 84.94% — Imported 2026-05-27
2 anthropic-cua 81.77% — Imported 2026-05-27
3 gemini-3.1 73.39% — Imported 2026-05-27
4 qwen-3 58.38% — Imported 2026-05-27
5 kimi-k2-5 55.98% — Imported 2026-05-27
6 claude-opus-4-6 49.24% — Imported 2026-05-27
7 gpt-5.4 43.49% — Imported 2026-05-27