Switchcraft: AI Model Router for Agentic Tool Calling
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866913101479149568 |
|---|---|
| author | Agarwal, Sharad Namyar, Pooria Wolman, Alec Ambavat, Rahul Gupta, Ankur Zhang, Qizheng |
| author_facet | Agarwal, Sharad Namyar, Pooria Wolman, Alec Ambavat, Rahul Gupta, Ankur Zhang, Qizheng |
| contents | Agentic AI systems that invoke external tools are powerful but costly, leading developers to default to large models and overspend inference budgets. Model routing can mitigate this, but existing routers are designed for chat completion rather than tool use. We present Switchcraft, the first (to the best of our knowledge) model router optimized for agentic tool calling. Switchcraft operates inline, selecting the lowest-cost model subject to correctness. We construct an evaluation framework on five function-calling benchmarks and train a DistilBERT-based classifier, deployed under a latency budget. Switchcraft achieves 82.9% accuracy -- matching or exceeding the best individual model -- while reducing inference cost by 84%, saving over $3,600 per million queries. We find that larger models do not consistently outperform smaller ones on tool-use tasks, and that nominally cheaper models can incur higher total cost due to token-intensive reasoning. Our work enables cost-aware agentic AI deployment without sacrificing correctness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_07112 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Switchcraft: AI Model Router for Agentic Tool Calling Agarwal, Sharad Namyar, Pooria Wolman, Alec Ambavat, Rahul Gupta, Ankur Zhang, Qizheng Artificial Intelligence Multiagent Systems Agentic AI systems that invoke external tools are powerful but costly, leading developers to default to large models and overspend inference budgets. Model routing can mitigate this, but existing routers are designed for chat completion rather than tool use. We present Switchcraft, the first (to the best of our knowledge) model router optimized for agentic tool calling. Switchcraft operates inline, selecting the lowest-cost model subject to correctness. We construct an evaluation framework on five function-calling benchmarks and train a DistilBERT-based classifier, deployed under a latency budget. Switchcraft achieves 82.9% accuracy -- matching or exceeding the best individual model -- while reducing inference cost by 84%, saving over $3,600 per million queries. We find that larger models do not consistently outperform smaller ones on tool-use tasks, and that nominally cheaper models can incur higher total cost due to token-intensive reasoning. Our work enables cost-aware agentic AI deployment without sacrificing correctness. |
| title | Switchcraft: AI Model Router for Agentic Tool Calling |
| topic | Artificial Intelligence Multiagent Systems |
| url | https://arxiv.org/abs/2605.07112 |