---
title: "Get Benchmark Task Outcomes"
method: GET
path: "/api/v1/benchmarks/{bench_id}/task-outcomes"
tags: ["benchmarks"]
---

# Get Benchmark Task Outcomes

`GET /api/v1/benchmarks/{bench_id}/task-outcomes`

## Request

- Base URL: `https://api.trajectory.ai` (the document declares no server; this is the origin it was published from)
- URL: `https://api.trajectory.ai/api/v1/benchmarks/{bench_id}/task-outcomes`
- Auth: none declared

Task outcomes across every valid training and evaluation run of this benchmark, independent of task split. Counts and mean rewards exclude execution errors from completed outcomes; ungraded completed attempts remain unscored.

## Path parameters

- `bench_id` string, required

## Query parameters

- `task_id` string[], required
- `include_recent_trajectories` boolean: Include each task's 10 most recent trajectories, newest first, and every trajectory counted by outcome.
- `run_id` union: Count only these training or evaluation runs' trajectories.
  - string[]
  - string

## Response `200`

Successful Response

- BenchmarkTaskOutcomes
  - `items` BenchmarkTaskOutcome[], required
    - `task_id` string, required
    - `total_count` integer, required
    - `completed_count` integer, required
    - `error_count` integer, required
    - `ungraded_count` integer, required
    - `average_reward` number, nullable, required
    - `recent_trajectories` BenchmarkTaskRecentTrajectory[], nullable: The task's most recent trajectories, newest first; present only when requested.
      - `trajectory_id` string, required
      - `created_at` string, date-time, required
      - `run_source` 'training' | 'eval', required
      - `run_id` string, required
      - `run_name` string, nullable, required
      - `status` string, nullable, required
      - `termination_reason` string, nullable, required
      - `reward` number, nullable, required
    - `outcome_breakdown` BenchmarkTaskOutcomeBreakdown
      - `passed` integer, required
      - `partial` integer, required
      - `failed` integer, required
      - `error` integer, required
      - `cancelled` integer, required
      - `running` integer, required
      - `ungraded` integer, required

## Changes

- **2026-10-31** `7b218d495d5c`: 1 info
  - `query` request parameter `run_id` list-of-types was widened by adding types `array`
- **2026-10-23** `a99a7b2a5444`: 4 info
  - added the new optional `query` request parameter `include_recent_trajectories`
  - added the new optional `query` request parameter `run_id`
  - added the optional property `items/items/outcome_breakdown` to the response with the `200` status
  - added the optional property `items/items/recent_trajectories` to the response with the `200` status
- **2026-10-16** `9213d826e063`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/trajectory/apis/trajectory-api/changes/api/v1/benchmarks/:bench_id/task-outcomes/get.md)

---

[API](https://skmtc.dev/trajectory/apis/trajectory-api.md) - [All operations](https://skmtc.dev/trajectory/apis/trajectory-api/llms.txt) - [OpenAPI document](https://skmtc.dev/trajectory/apis/trajectory-api/revisions/5b5a8ae4b4a4?raw)