---
title: "Start Benchmark Evaluation"
method: POST
path: "/felix/benchmarks/evaluate"
tags: ["felix"]
---

# Start Benchmark Evaluation

`POST /felix/benchmarks/evaluate`

Start an asynchronous benchmark evaluation.

Launches evaluation in the background and returns immediately with
an evaluation_id that can be used to check status.

## Request body

- BenchmarkEvaluationRequest — Request to run a benchmark evaluation
  - `model_id` string, required — Training job ID, explicit GLiNER base model ID ('fastino/gliner2-base-v1'), or decoder base model ID ('base:{hf_model_id}')
  - `project_id` string, nullable — Optional project ID to associate the benchmark evaluation with
  - `task` 'ner' | 'text_classification', required — Task type for evaluation
  - `benchmark` string, required — Benchmark name (e.g., 'fewnerd', 'mteb/banking77', 'tweeteval')
  - `max_samples` integer — Maximum samples to evaluate
  - `split` string — Dataset split to use
  - `benchmark_config` object, nullable — Additional benchmark-specific configuration

## Response `200`

Successful Response

- BenchmarkEvaluationResponse — Response for benchmark evaluation
  - `success` boolean, required
  - `evaluation_id` string, required
  - `project_id` string, nullable
  - `status` 'pending' | 'running' | 'complete' | 'failed' | 'cancelled', required
  - `task` string, required
  - `benchmark` string, required
  - `model_id` string, required
  - `metrics` object, nullable
  - `error_message` string, nullable
  - `created_at` string, date-time, required
  - `completed_at` string, date-time, nullable

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.net/pioneer/apis/brain-api.md) · [All operations](https://skmtc.net/pioneer/apis/brain-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/pioneer/brain-api/versions/31dfe831e079/schema)