# Tool-use Agent Scalability and Performance on Tau2-Bench

Source: [Tool-use Agent Scalability and Performance on Tau2-Bench](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-21T07:14:36.604Z

Evaluation focusing on the operational efficiency and scalability of tool-using agents, measured by runtime, trajectory count, and token throughput on the Tau2-Bench dataset.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Tool-use agent scalability and performance |
| Dataset (including any stated split) | Tau2-Bench |
| Metric | Runtime |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 2 of 2 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Runtime](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench.md?metric=Runtime>): Lower is better; unit: TIME.

- [Total Trajectories](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench.md?metric=Total%20Trajectories>): Higher is better; unit: SCALAR.

- [Average Reward](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench.md?metric=Average%20Reward>): Higher is better; unit: SCALAR.

- [Tokens Generated](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench.md?metric=Tokens%20Generated>): Higher is better; unit: SCALAR.

- [Tokens per Second](<https://www.sota2.com/research/sota/tool-use-agent-scalability-and-performance-on-tau2-bench.md?metric=Tokens%20per%20Second>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Runtime | Result date | Report paper |
| --- | --- | --- | --- |
| Tau2-Agent | 1 | Not recorded | 6927db58d6b0dd3342930f9d |
| P2P-Agent | 1 | Not recorded | 6927db58d6b0dd3342930f9d |

### Tau2-Agent

Runtime: 1. Result ID: 69e7228947f34ce924ccdb66.

[Reporting paper](<https://www.sota2.com/research/paper/6927db58d6b0dd3342930f9d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa79391e209131cea7cb4a.md>)

Recorded evaluation configuration: {"Concurrent tasks":"500"}. This record does not assert independent reproduction.

### P2P-Agent

Runtime: 1. Result ID: 69e7228947f34ce924ccdb67.

[Reporting paper](<https://www.sota2.com/research/paper/6927db58d6b0dd3342930f9d.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db58d6b0dd3342930f9d.md>)

Recorded evaluation configuration: {"Concurrent tasks":"1,500"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.