Published August 2, 2026
| Version v1.7.0
Software
Open
AWB: AI Workflow Benchmark
Authors/Creators
Description
AWB is an open benchmark harness that measures AI coding tool + workflow performance on 100 real-repository tasks pinned to commit SHAs. It scores seven sigmoid-normalized dimensions (correctness, cost, speed, code quality, reliability, security, efficiency) and reports a Production Readiness Score plus a Workflow Lift between vanilla and configured tool runs. Trace artifacts use OpenTelemetry GenAI semantic conventions.
Notes
If you use AWB in research, please cite this entry.
Files
xmpuspus/ai-workflow-benchmark-v1.7.0.zip
Files
(1.7 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:9763a282f7845ee635564328e436d5e7
|
1.7 MB | Preview Download |
Additional details
Related works
- Is supplement to
- Software: https://github.com/xmpuspus/ai-workflow-benchmark/tree/v1.7.0 (URL)
Software
- Repository URL
- https://github.com/xmpuspus/ai-workflow-benchmark