Get Evaluation Results

Retrieve the complete results of a finished evaluation.

This endpoint returns detailed evaluation metrics and scores for a completed evaluation. Works for both single model evaluations and comparison mode evaluations.

Path Parameters:

• evaluation_id: Unique evaluation identifier

Returns:

• evaluation_id: The evaluation identifier
• model_id: Primary model that was evaluated
• benchmark_id: Benchmark that was used
• status: Evaluation status (should be completed)
• raw_answers_count: Number of raw answers generated during evaluation
• completed_at: ISO timestamp when evaluation finished
• method (optional): Evaluation method (eval for single model, eval-compare for comparison)
• metrics (optional): Evaluation metrics and scores (single model only)
• model_id_2 (optional): Second model ID (comparison mode only)
• base_model (optional): Base model results (comparison mode only)
• eval_model (optional): Eval model results (comparison mode only)
• comparison (optional): Comparison results between models (comparison mode only)

Raises:

• 404: If evaluation not found or doesn’t belong to the authenticated user
• 400: If evaluation is not yet completed

Example Request:

GET /api/v3/evaluations/eval-xyz789/results
Headers: {"Authorization": "Bearer <api_key>"}

Example Response (Single Model):

{
  "evaluation_id": "eval-xyz789",
  "model_id": "nugen-flash-instruct",
  "benchmark_id": "task-abc123",
  "status": "completed",
  "method": "eval",
  "raw_answers_count": 10,
  "completed_at": "2024-01-15T10:45:00Z",
  "metrics": {
    "accuracy": 0.92,
    "relevance": 0.88,
    "average_score": 0.90,
    "total_questions": 10,
    "correct_answers": 9
  }
}

Example Response (Comparison Mode):

{
  "evaluation_id": "eval-xyz789",
  "model_id": "nugen-flash-instruct",
  "model_id_2": "gpt-4",
  "benchmark_id": "task-abc123",
  "status": "completed",
  "method": "eval-compare",
  "raw_answers_count": 20,
  "completed_at": "2024-01-15T10:45:00Z",
  "base_model": {
    "model_id": "nugen-flash-instruct",
    "average_score": 0.92,
    "total_questions": 10
  },
  "eval_model": {
    "model_id": "gpt-4",
    "average_score": 0.85,
    "total_questions": 10
  },
  "comparison": {
    "winner": "nugen-flash-instruct",
    "score_difference": 0.07,
    "statistical_significance": true
  }
}

Notes:

• Only available for completed evaluations
• Single model evaluations use the metrics field for results
• Comparison mode evaluations use base_model, eval_model, and comparison fields instead
• The method field indicates whether it’s a single model (eval) or comparison (eval-compare) evaluation
• Use /evaluations/{evaluation_id}/status first to check if evaluation is complete
• For downloadable format, use /evaluations/{evaluation_id}/download