evals
Create Run
Create a new run for an evaluation.
Model-based routing (for multi-account support):
- Pass model via header: x-litellm-model: gpt-4-account-1
- Pass model via query: ?model=gpt-4-account-1
- Pass model via body: {"model": "gpt-4-account-1"}
- Pass model via completion.model: {"completion": {"model": "gpt-4-account-1"}}
Example usage:
curl -X POST "http://localhost:4000/v1/evals/eval_123/runs" -H "Authorization: Bearer your-key" -H "Content-Type: application/json" -d '{
"data_source": {"type": "dataset", "dataset_id": "dataset_123"},
"completion": {"model": "gpt-4", "temperature": 0.7}
}'
Returns: Run object with id, status, timestamps, etc.
post/v1/evals/{eval_id}/runs
Path parameters
eval_idstring required
Query parameters
custom_llm_providerstring nullable
Response
Successful Response
Changes
Changed in 2 of the 42 revisions of this API.6
- ○
api tag
evalsaddedapi-tag-added
- ○
api tag
OpenAI Evals API - Runsremovedapi-tag-removed
- ○
- ○
api tag
OpenAI Evals API - Runsaddedapi-tag-added
- ○
api tag
evalsremovedapi-tag-removed
- ○
added the optional property
//to the response with the statusresponse-optional-property-added
- ○
added the optional property
//to the response with the statusresponse-optional-property-added
This revision also has 5 changes that name no endpoint, such as unreferenced schemas being removed. See the revision's changelog
- ○