ArchBench Microservice Generation

Microservice Generation: Generate complete microservice implementations from requirements and codebase context.

3rows
clean_state_test_pass_rate_p1primary metric
2026-05-06sampled

Metadata

Metrics

Incremental Test Pass Rate P1, Incremental Test Pass Rate P2, Clean-State Test Pass Rate P1, Clean-State Test Pass Rate P2, Average Time (lower is better), Average Cost (lower is better), Average Input Tokens (lower is better), Average Output Tokens (lower is better)

Latest Results

Rows ranked by clean_state_test_pass_rate_p1. ArchBench paper: https://arxiv.org/abs/2603.17833. CLI: https://github.com/sa4s-serc/archbench-cli.

Rank Subject Clean-State Test Pass Rate P1 Model Match Provenance Sampled
1 Codex 98.10 Imported 2026-05-06
2 Claude Code 96.90 Imported 2026-05-06
3 Code Qwen 81.90 Imported 2026-05-06