PinchBench
Agents · 2026-02-10
PinchBench evaluates OpenClaw agents on real-world tool-use tasks. This root represents source-reported overall results; qualified cross-run aggregations such as a model's best observed score use child metrics.
Top models (higher is better)
| Model | Score |
|---|---|
| Nemotron 3.5 Bolt 74B A9B BF16 | 90.3 |
| Nemotron 3 Ultra 550B A55B | 90.0 |
| Nemotron 3 Ultra 550B A55B (NVFP4) | 89.8 |