# Long-horizon Agentic Reasoning on GAIA

Source: [Long-horizon Agentic Reasoning on GAIA](<https://www.sota2.com/research/sota/long-horizon-agentic-reasoning-on-gaia>)

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

Data updated: 2026-06-18T08:04:25.739Z

Evaluation of Long-horizon Agentic Reasoning performance on the GAIA dataset, measured by Pass@3.

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 | Long-horizon Agentic Reasoning |
| Dataset (including any stated split) | GAIA |
| Metric | Pass@3 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 3 of 3 |
| 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

- [Pass@3](<https://www.sota2.com/research/sota/long-horizon-agentic-reasoning-on-gaia.md?metric=Pass%403>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Pass@3 | Result date | Report paper |
| --- | --- | --- | --- |
| AgentCPM-Explore + Long-context RL | 70.9 | Not recorded | 6a337bc8206542251d7c523e |
| AgentCPM-Explore + Long-context RL | 67.7 | Not recorded | 6a337bc8206542251d7c523e |
| AgentCPM-Explore | 66.1 | Not recorded | 6a337bc8206542251d7c523e |

### AgentCPM-Explore + Long-context RL

Pass@3: 70.9. Result ID: 6a33a67ffbc373592c30c77f.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bc8206542251d7c523e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bc8206542251d7c523e.md>)

Recorded evaluation configuration: {"Training Steps":"50 steps","Data Recipe":"Long-context data recipe"}. This record does not assert independent reproduction.

### AgentCPM-Explore + Long-context RL

Pass@3: 67.7. Result ID: 6a33a67ffbc373592c30c77e.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bc8206542251d7c523e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6a337bc8206542251d7c523e.md>)

Recorded evaluation configuration: {"Training Steps":"25 steps","Data Recipe":"Long-context data recipe"}. This record does not assert independent reproduction.

### AgentCPM-Explore

Pass@3: 66.1. Result ID: 6a33a67ffbc373592c30c77d.

[Reporting paper](<https://www.sota2.com/research/paper/6a337bc8206542251d7c523e.md>)

[Method origin paper](<https://www.sota2.com/research/paper/698969e4c97a0b3e3b5dc3a5.md>)

Recorded evaluation configuration: {"Data Recipe":"Base Agent-enhanced model"}. This record does not assert independent reproduction.

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