evals

Get detailed information about a specific evaluation.

get/api/v1/evaluations/{evaluation_id}

Request

  • Base URL
    https://api.primeintellect.ai
  • Auth
    HTTP bearer

Path parameters

evaluation_idstring required

Response

Successful Response

evaluation_idstring required
namestring required
status'PENDING' | 'RUNNING' | 'PROCESSING' | 'COMPLETED' | 'FAILED' | 'TIMEOUT' | 'CANCELLED' required

Evaluation status enum

eval_type'suite' | 'training' | 'environment' required

Evaluation type enum

user_idstring nullable

User ID of evaluation owner

team_idstring nullable

Team ID if evaluation is owned by a team

environment_idsstring[] nullable
environment_namesstring[] nullable
suite_idstring nullable
run_idstring nullable
version_idstring nullable
has_rollout_errorsboolean
is_hostedboolean
sandbox_idstring nullable
inference_modelstring nullable
eval_configobject nullable
error_messagestring nullable
logsstring nullable
model_namestring nullable
datasetstring nullable
frameworkstring nullable
task_typestring nullable
descriptionstring nullable
tagsstring[]
total_samplesinteger required
avg_scorenumber nullable
min_scorenumber nullable
max_scorenumber nullable
started_atstring date-time nullable
completed_atstring date-time nullable
created_atstring date-time required
updated_atstring date-time nullable
metadataobject nullable
metricsobject nullable
viewer_urlstring nullable
is_publicboolean

Whether this evaluation is publicly shareable by link

show_on_leaderboardboolean

Whether this public evaluation appears on environment leaderboards

statisticsobject nullable

Statistics format (totalResults, totalReward, avgReward)

chartDataobject nullable

Histogram and statistics for visualization

rubricInfoobject nullable

Rubric with detected metrics and weights

detailedMetricsobject nullable

Detailed statistics for each detected metric

rewardStatsobject nullable

Reward statistics for backward compatibility

Changes

    • breaking

      the response property became optional for the status

    • breaking

      response property updated_at list-of-types was widened by adding types null to media type application/json of response 200

    • info

      added the optional property to the response with the status