LoopTool-8B: benchmark results
Provider: Other.
Aggregate Rating 1563 ± 50, rank #1505 of 10916 rated models, from 24 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| RobustBench-TC - Misleading Description (Neutral Suffix) | 48 | Tool-call accuracy (%) under the MisDesc-N perturbation (rew | 88.2 |
| RobustBench-TC - Timeout | 69.8 | Tool-call accuracy (%) under the Timeout perturbation (trans | 88.2 |
| RobustBench-TC - Clean | 71.4 | Tool-call accuracy (%) on the 199 unperturbed baseline sampl | 82.4 |
| RobustBench-TC - Malformed Response | 65.8 | Tool-call accuracy (%) under the Malform perturbation (trans | 82.4 |
| RobustBench-TC - Query Paraphrase | 68.8 | Tool-call accuracy (%) under the QueryPara perturbation (obs | 82.4 |
| RobustBench-TC - Duplicate Tool Without Description | 52.6 | Tool-call accuracy (%) under the Dup-NoDesc perturbation (ac | 76.5 |
| RobustBench-TC - Misleading Description | 37.3 | Tool-call accuracy (%) under the MisDesc perturbation (rewar | 76.5 |
| RobustBench-TC - Parameter Description Paraphrase | 68.8 | Tool-call accuracy (%) under the ParamPara perturbation (obs | 76.5 |
| RobustBench-TC - Redundant Tool | 59.3 | Tool-call accuracy (%) under the RedunTool perturbation (act | 73.5 |
| RobustBench-TC - Tool Description Paraphrase | 69.8 | Tool-call accuracy (%) under the ToolPara perturbation (obse | 73.5 |
| RobustBench-TC - Rate Limit | 27.6 | Tool-call accuracy (%) under the RateLim perturbation (trans | 70.6 |
| RobustBench-TC - Duplicate Tool With Description and Wrong Parameters | 52.8 | Tool-call accuracy (%) under the Dup-DescWP perturbation (ac | 67.6 |
Interactive version: theaggregate.ai/model?slug=looptool-8b · How It Works · Data refreshed daily, snapshot 2026-11-08.