---
title: "Get all evaluation jobs"
method: GET
path: "/v1/evaluation"
tags: ["evaluation"]
---

# Get all evaluation jobs

`GET /v1/evaluation`

## Request

- Base URL: `https://api.together.ai`
- URL: `https://api.together.ai/v1/evaluation`
- Auth: HTTP bearer

## Query parameters

- `status` string: Filter evaluation jobs by status
- `limit` integer: Limit the number of results

## Response `200`

evaluation jobs retrieved successfully

- EvaluationJob[]
  - `workflow_id` string: The evaluation job ID
  - `type` 'classify' | 'score' | 'compare': The type of evaluation
  - `owner_id` string: ID of the job owner (admin only)
  - `status` 'pending' | 'queued' | 'running' | 'completed' | 'error' | 'user_error': Current status of the job
  - `status_updates` EvaluationJobStatusUpdate[]: History of status updates (admin only)
    - `status` string: The status at this update
    - `message` string: Additional message for this update
    - `timestamp` string, date-time: When this update occurred
  - `parameters` object: The parameters used for this evaluation
  - `created_at` string, date-time: When the job was created
  - `updated_at` string, date-time: When the job was last updated
  - `results` union: Results of the evaluation (when completed)
    - EvaluationClassifyResults
      - `generation_fail_count` number, nullable: Number of failed generations.
      - `judge_fail_count` number, nullable: Number of failed judge generations
      - `invalid_label_count` number, nullable: Number of invalid labels
      - `result_file_id` string: Data File ID
      - `pass_percentage` number, nullable: Pecentage of pass labels.
      - `label_counts` string: JSON string representing label counts
    - EvaluationScoreResults
      - `aggregated_scores` object
        - `mean_score` number
        - `std_score` number
        - `pass_percentage` number
      - `generation_fail_count` number, nullable: Number of failed generations.
      - `judge_fail_count` number, nullable: Number of failed judge generations
      - `invalid_score_count` number: number of invalid scores generated from model
      - `failed_samples` number: number of failed samples generated from model
      - `result_file_id` string: Data File ID
    - EvaluationCompareResults
      - `A_wins` integer: Number of samples where model A was judged the winner
      - `B_wins` integer: Number of samples where model B was judged the winner
      - `Ties` integer: Number of samples that resulted in a tie score
      - `generation_fail_count` number, nullable: Number of generation failures across model A and model B.
      - `judge_fail_count` number, nullable: Number of judge inference failures. In the default two-pass mode (disable_position_bias_correction=false) this is the combined failure count from both the original-order and flipped-order judge passes.
      - `result_file_id` string: File ID of the detailed output file. Each row contains the original input fields plus judge outputs. In two-pass mode the file includes both original-order and flipped-order judge fields; in single-pass mode (disable_position_bias_correction=true) only original-order fields are present.
    - object
      - `error` string

## Other responses

- `400`: Invalid request format
- `500`: Error retrieving jobs from manager

## Changes

- **2026-10-22** `bb5257b3173a`: 1 info
  - endpoint added

[Operation history](https://skmtc.dev/together/apis/together-apis/changes/v1/evaluation/get.md)

---

[API](https://skmtc.dev/together/apis/together-apis.md) - [All operations](https://skmtc.dev/together/apis/together-apis/llms.txt) - [OpenAPI document](https://skmtc.dev/together/apis/together-apis/revisions/93438ddac711?raw)