Capability
Agentic and Tool-Use Benchmarks
Agentic and Tool-Use Benchmarks. Demo series compiled by The Latent research desk.
Agentic and Tool-Use Benchmarks
- SWE-bench Verified
- OSWorld
- GAIA
View data
| Date | SWE-bench Verified | OSWorld | GAIA |
|---|---|---|---|
| 2026-01-01 | 72.41 | 47.87 | 64.2 |
| 2026-02-01 | 79.96 | 52.83 | 70.82 |
| 2026-03-01 | 82.04 | 54.26 | 72.8 |
| 2026-04-01 | 82.23 | 54.36 | 72.91 |
| 2026-05-01 | 85.29 | 56.42 | 75.71 |
| 2026-06-01 | 78.59 | 52.02 | 69.85 |
| 2026-07-01 | 77.6 | 51.3 | 68.8 |
API access · Pro coming soon
Demo preview through July 2026 (percent).
Methodology
Synthetic demo series generated to preview the data product layout.