> ## Documentation Index
> Fetch the complete documentation index at: https://agent-compass.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# SWE-bench Pro

> Harder professional repository tasks.

SWE-bench Pro evaluates harder professional repository tasks and uses `/workspace` as the default remote root.

## Runtime Status

| Field               | Value                                                        |
| ------------------- | ------------------------------------------------------------ |
| Benchmark id        | `swebench_pro`                                               |
| Tags                | `Agentic Coding`, `Repository Repair`, `Long-Horizon`        |
| Execution type      | remote environment                                           |
| Typical harness     | `mini_swe_agent, openhands, claude_code, codex, or openclaw` |
| Typical environment | `docker, daytona, or modal`                                  |
| Current status      | registered in the direct runtime                             |

## When to Use

Use SWE-bench Pro when you need to measure agentic coding behavior with the task assumptions described by this benchmark. For large or remote benchmarks, prefer benchmark recipes so images, workspaces, and provider-specific defaults come from task metadata instead of manual CLI flags.

## Parameters

Common parameters for this benchmark include:

* `sample_ids`
* `split`
* `k`
* `avgk`
* `resume`

Shared runtime controls such as `k`, `avgk`, `sample_ids`, `resume`, and `category` follow the conventions in [Benchmark Parameters](/reference/benchmarks/overview).

## Run Example

```bash theme={null}
agentcompass run \
  swebench_pro \
  mini_swe_agent \
  your-model \
  --env <env-provider> \
  --benchmark-params '{"sample_ids":["<task-id>"]}' \
  --model-base-url "$MODEL_BASE_URL" \
  --model-api-key "$MODEL_API_KEY" \
  --model-api-protocol openai-chat
```

Adjust the harness and environment to the supported combination for your branch and deployment.

## Outputs

Per-task details are written to `results/swebench_pro/<model>/<run>/details/`. Aggregate metrics are written to `summary.md` in the same run directory.

## Notes

Use provider recipes so task-specific dependencies and workspaces are prepared before the harness starts.