Get Eval Run

GET

/org/:org_id/project/:project_id/task/:task_id/eval/run/:eval_run_id

GET

/api/v1/org/:org_id/project/:project_id/task/:task_id/eval/run/:eval_run_id

$ curl https://app.rightbrain.ai/api/v1/org/org_id/project/project_id/task/task_id/eval/run/eval_run_id \
>      -H "Authorization: Bearer <token>"

1 {
2   "id": "019fa4ed-4fec-f5c1-e25a-7dc65ada66ca",
3   "task_id": "019f8a00-1c3d-7e42-b8f1-2a5c9d4e7b60",
4   "project_id": "019f6b10-2e8a-7c91-9a3d-6e0f8b2c1d47",
5   "candidate_revision_id": "019f8a06-aa82-7da7-b159-8bb1d36bab06",
6   "task_run_ids": [
7     "019f8a04-8e60-7b85-af37-6e9fb14d8fa4",
8     "019f8a05-9f71-7c96-b048-7fa0c25f9ab5"
9   ],
10   "status": "completed",
11   "pass_count": 2,
12   "fail_count": 0,
13   "error_count": 0,
14   "created": "2026-07-27T18:54:03.589524Z",
15   "modified": "2026-07-27T18:54:04.601868Z",
16   "judge_model": {
17     "name": "claude-sonnet-5",
18     "provider": "Anthropic"
19   },
20   "pass_rate": 1,
21   "eval_set_id": "019fa4ec-98a8-9c34-e8f6-09f54481dd89",
22   "started_at": "2026-07-27T18:54:04.584779Z",
23   "completed_at": "2026-07-27T18:54:11.458892Z",
24   "created_by_user_id": "019f6c10-1a2b-7c3d-8e4f-5a6b7c8d9e01",
25   "metrics": {
26     "token_cost": {
27       "reference_avg": 215,
28       "candidate_avg": 239,
29       "delta_pct": 11.16
30     },
31     "charged_credits": {
32       "reference_avg": 1,
33       "candidate_avg": 1,
34       "delta_pct": 0
35     },
36     "llm_call_timing": {
37       "reference_avg": 1.21,
38       "candidate_avg": 2.11,
39       "delta_pct": 74.35
40     }
41   },
42   "results": [
43     {
44       "id": "019fa4ed-6bda-b54e-bc1d-d24d82d368e5",
45       "eval_run_id": "019fa4ed-4fec-f5c1-e25a-7dc65ada66ca",
46       "reference_run_id": "019f8a04-8e60-7b85-af37-6e9fb14d8fa4",
47       "judge_input_tokens": 1026,
48       "judge_output_tokens": 232,
49       "is_error": false,
50       "created": "2026-07-27T18:54:10.912400Z",
51       "candidate_run_id": "019fa4ed-5ffe-11e3-5898-7985979a6961",
52       "verdict": "pass",
53       "reasoning": "The candidate includes all fields from the reference (summary, category, priority) with identical values, plus an additional \"churn_risk\": false field. Since \"churn_risk\" is part of the expected output schema, its inclusion is appropriate rather than a hallucination.",
54       "metrics": {
55         "reference": {
56           "total_tokens": 215,
57           "charged_credits": "1.00",
58           "llm_call_timing": 1.37
59         },
60         "candidate": {
61           "total_tokens": 239,
62           "charged_credits": "1.00",
63           "llm_call_timing": 3.1
64         }
65       }
66     },
67     {
68       "id": "019fa4ed-6da5-5941-4c3e-07c0831947b0",
69       "eval_run_id": "019fa4ed-4fec-f5c1-e25a-7dc65ada66ca",
70       "reference_run_id": "019f8a05-9f71-7c96-b048-7fa0c25f9ab5",
71       "judge_input_tokens": 1051,
72       "judge_output_tokens": 250,
73       "is_error": false,
74       "created": "2026-07-27T18:54:11.371402Z",
75       "candidate_run_id": "019fa4ed-585c-d185-be13-5659b377377e",
76       "verdict": "pass",
77       "reasoning": "The candidate output includes all fields from the reference (summary, category, priority) with identical values, and additionally includes \"churn_risk\": false. This is a schema-defined field the reference simply omitted, not a hallucination.",
78       "metrics": {
79         "reference": {
80           "total_tokens": 215,
81           "charged_credits": "1.00",
82           "llm_call_timing": 1.05
83         },
84         "candidate": {
85           "total_tokens": 239,
86           "charged_credits": "1.00",
87           "llm_call_timing": 1.11
88         }
89       }
90     }
91   ]
92 }

Get eval run details with results and aggregate metrics.

In addition to the fields on EvalRunResponse, this returns:

judge_model: the model used to produce verdicts
metrics: reference vs. candidate averages (token cost, charged credits, LLM call timing) with the percentage delta between them
results: one entry per reference/candidate run pair, each with the judge’s verdict (pass/fail), free-text reasoning, and per-pair token and cost metrics

metrics and results reflect the state at the time of the request; while status is pending or running, results may be a partial or empty list.

Get eval run details with results and aggregate metrics. In addition to the fields on `EvalRunResponse`, this returns: - `judge_model`: the model used to produce verdicts - `metrics`: reference vs. candidate averages (token cost, charged credits, LLM call timing) with the percentage delta between them - `results`: one entry per reference/candidate run pair, each with the judge's `verdict` (`pass`/`fail`), free-text `reasoning`, and per-pair token and cost metrics `metrics` and `results` reflect the state at the time of the request; while `status` is `pending` or `running`, `results` may be a partial or empty list.

Authentication

AuthorizationBearer

API key or access token, sent as Authorization: Bearer <token>. The simplest option: create an API key in your dashboard under Settings, then paste it here.

AuthorizationBearer

API key or access token, sent as Authorization: Bearer <token>. The simplest option: create an API key in your dashboard under Settings, then paste it here.

Path parameters

eval_run_idstringRequiredformat: "uuid"

org_idstringRequiredformat: "uuid"

The organization ID

project_idstringRequiredformat: "uuid"

The project ID

task_idstringRequiredformat: "uuid"

The task ID

Response

Eval run detail retrieved successfully

idstringformat: "uuid"

task_idstringformat: "uuid"

project_idstringformat: "uuid"

candidate_revision_idstringformat: "uuid"

task_run_idslist of strings

statusenum

pass_countinteger

fail_countinteger

error_countinteger

createdstring

modifiedstring

judge_modelobject

pass_ratedouble or nullRead-only

eval_set_idstring or nullOptionalformat: "uuid"

started_atstring or nullOptional

completed_atstring or nullOptional

error_messagestring or nullOptional

created_by_user_idstring or nullOptionalformat: "uuid"

metricsobject or nullOptional

resultslist of objectsOptionalDefaults to []

Errors

400

Bad Request Error

401

Unauthorized Error

403

Forbidden Error

404

Not Found Error

422

Unprocessable Entity Error