LIVE
The AI Parity Index0.1%Tasks measured1 of 18,796Unemployment4.1% · Aug 2026Payrolls+162K m/m · Aug 2026159.1M employedJob openings7.1M · Aug 20260.3M vs Jul 2026Wages$37.75/h avg · Aug 2026+3.1% y/yProductivity+1.4% q/q ann. · Q2 2026Inflation+3.4% y/y · Aug 2026Next upSoftware Developers · 16 more tasksNext releaseQ4 2026
The AI Parity Index0.1%Tasks measured1 of 18,796Unemployment4.1% · Aug 2026Payrolls+162K m/m · Aug 2026159.1M employedJob openings7.1M · Aug 20260.3M vs Jul 2026Wages$37.75/h avg · Aug 2026+3.1% y/yProductivity+1.4% q/q ann. · Q2 2026Inflation+3.4% y/y · Aug 2026Next upSoftware Developers · 16 more tasksNext releaseQ4 2026
Modify existing software to correct errors, adapt it to new hardware, or upgrade interfaces and improve performance.
1.7M jobs · $71/h avg · $46B of GDP a year · 6.2% of the job · verification method: machine-checked
Leaderboard
The coding agents people use every day, on the same ticket. The measured loop is our minimal reference harness: test, patch, retry.
| # | Harnessranked by success | Success rateshare of runs at the bar | Consistentruns passed | Cost per tasklist-price est. | Time per taskinference | vs human× output · − pts |
|---|---|---|---|---|---|---|
| Human baseline | 100% | — | ~$225 est. | ~25 min est. | — | |
| 1 |  Test-and-retry loop · Claude Opus › | 93% | 25/27 | $0.29 | 19s | -7 pts |
| · |  Claude Code | invited · not yet run | ||||
| · |  Cursor | invited · not yet run | ||||
| · |  OpenAI Codex | invited · not yet run | ||||
| · |  Gemini CLI | invited · not yet run | ||||
| · |  Devin | invited · not yet run | ||||
| Verification method | Machine-checked · rubric v1.0, model-applied, not yet human-reviewed |
| Industries | Professional and Technical Services 41% · Information 21% · Finance and Insurance 10% · Manufacturing 8% |
| Source | O*NET task 21670 · Core · importance 3.74 of 5 · jobs and wages BLS OEWS May 2025 |
Working prototype by Recursiv Labs · not an official SPARK AI or UC San Diego publication · Data · © 2026