Chi-Bench: Long-Horizon Healthcare Workflow Agent Benchmark

χ-Bench (Chi-Bench) is a benchmark dataset for evaluating AI agents on end-to-end U.S. healthcare workflows. It was created by author 'actava' and last updated on 2026-05-19. The dataset provides task fixtures across three long-horizon domains: provider prior authorization, payer utilization management, and population care management.

Use Cases

• Benchmarking AI agent performance on provider prior authorization workflows based on the described clinical case tasks.
• Evaluating agent decision-making in payer utilization management scenarios based on the described policy-rich environment.
• Testing long-horizon planning capabilities for population care management based on the described end-to-end workflow simulation.

Strengths

• Focuses on three specific, complex healthcare domains: prior authorization, utilization management, and care management.
• Simulates a high-fidelity environment of 20 healthcare applications exposed over MCP.
• Benchmark tasks are designed for long-horizon, policy-rich agent evaluation.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-05-19 04:40:32; freshness should be verified.

Provenance

Source

actava on Hugging Face

Collection Method

Likely created as a benchmark task suite for AI agent evaluation.

Freshness

2026-05-19 04:40:32

Geography

U.S. healthcare workflows