Scale AI SEAL evaluation of LLM tool/function calling ability in enterprise scenarios with complex multi-step workflows.
Tests real enterprise tool-use patterns beyond simple function calls; expert evaluation by Scale AI.
Benchmark is deprecated by Scale AI; scores may not be updated with newer models.
Higher is better. Percentage of enterprise tool-use tasks completed correctly.
Benchmark is deprecated by Scale AI.
| # | Model | Score |
|---|---|---|
| 1 | 70.1 | |
| 2 | 68.8 | |
| 3 | 67.0 | |
| 4 | 66.4 | |
| 5 | 65.3 | |
| 6 | 65.3 | |
| 7 | 65.3 |