Skip to content
For the complete documentation index, see llms.txt. This page is also available as Markdown.

Evaluate Trained Models

Use Model Evaluation to explore how your model performs on your test dataset.

About

Model evaluations show:

  1. A production metrics explorer, which helps you find the optimal confidence threshold at which to run your model;

  2. Model improvement recommendations, which provide suggestions on how you can increase the accuracy of your model;

  3. Performance by class, which shows how well your model identifies different classes;

  4. A confusion matrix, which you can use to find specific classes on which your model thrives and struggles, and;

  5. An interactive vector explorer which lets you identify clusters of images where your model does well or poorly;

You can use model evaluation to identify areas of improvement for your model.

Model evaluations are automatically run for all versioned models trained on, or uploaded to Roboflow by paid users. It may take several minutes for an evaluation to run for a dataset of a few hundred images, and several hours for large datasets with thousands or more images.

The confusion matrix runs your trained model on the images in your test and validation sets and compares its predictions with your annotations (the ground truth). It shows correct predictions, false positives (the wrong class, or an object where none is present), and false negatives. The recommendations are based on it. They cover models with many false negatives or false positives, classes that are often confused, classes that need more data, and test or validation sets that may be too small.

Optimal Confidence

The production metrics explorer computes precision, recall, and F1 at every confidence threshold and recommends an optimal confidence: the threshold with the best trade-off. Once the evaluation completes, that threshold becomes the default for your model's inference requests. If per-class thresholds are available, those apply too, with the global threshold as the fallback for any class without its own value. To override it on a request, pass the confidence parameter.

Median Latency

The evaluation shows the median time the Serverless Cloud API takes to answer one request, measured on your test images. The value includes preprocessing and postprocessing time. It does not include network time. Your model lists and model cards show the same value as "Latency (Cloud API)". Latency from a Neural Architecture Search benchmark is measured a different way, so do not compare the two.

If your test images hold at least twice as many pixels as the model input, the evaluation adds a recommendation about capture resolution. The server decodes and resizes every image before it runs the model, so frames whose longer side is close to the model input answer faster.

Use Model Comparison to compare test-set accuracy and latency for models on the same Dataset Version.

Optimizing for Inference

After training, Roboflow compiles your model into the package the Serverless Cloud API serves. The model shows "Optimizing for inference" while this runs, and the evaluation waits for it to finish. This way the evaluation measures the same package that serves your inference requests. Compilation usually adds a few minutes. If it fails, or takes more than 24 hours, the evaluation starts anyway.

Supported Project Types

Model evaluation supports Object Detection, Instance Segmentation, Keypoint Detection, Classification, and Semantic Segmentation projects.

For Keypoint Detection, mAP uses Object Keypoint Similarity (OKS) to match predicted keypoints to labeled ones. The evaluation image viewer draws predicted keypoints, but not labeled keypoints.

For Semantic Segmentation, the headline metric is mIoU (mean Intersection-over-Union) instead of mAP. All metrics (precision, recall, F1) are computed at the pixel level rather than per-instance. The per-class breakdown shows IoU, precision, recall, F1, and an optimal confidence threshold for each class. Confusion matrix values represent pixel counts rather than object counts.

Newer evaluations of segmentation models also report boundary metrics. Boundary IoU compares narrow bands along each mask edge, so it shows edge errors that mask IoU can hide. Instance Segmentation shows a "Mean Boundary Average Precision" chart under "Show Detailed Metrics". Semantic Segmentation shows mBIoU next to mIoU, and a "Boundary IoU" column in the per-class table.

Web App

Open "Models" in your project, select a trained model, and click "View Full Model Evaluation".

In the production metrics explorer, drag the slider to see precision, recall, and F1 at any confidence threshold.

The model improvement recommendations list what to change to increase accuracy.

The performance by class chart counts correct predictions, misclassifications, false negatives, and false positives for each class. Use the "All Classes" dropdown to focus on specific classes.

In the confusion matrix, click a cell to see the images in it (ex: a cell in the "False Positive" column shows images where the model found an object your annotations do not have). Click an image, then switch between "Ground Truth" and "Model Predictions" to compare your annotations with the model's output.

"Vector Analysis" plots your test images by visual similarity and colors each by its F1 score. Click a point, an area, or one of the "Identified Clusters" to see those images, so you can find groups of similar images where the model struggles.

Vector Analysis with a selected cluster of two images highlighted on the scatter plot and listed by worst F1

The class chart and the confusion matrix start at the optimal confidence. Move their "Confidence Threshold" slider to see how the results, precision, and recall change.

Click "Improve My Model" to open a Roboflow Agent chat about this evaluation. The Agent shows the images where the model fails next to their labels and suggests changes to your data or training. It puts images with likely label errors into one review job, so you can check and fix the labels. It asks you before it moves images out of a teammate's job.

HTTP API

The Model Evaluations API returns everything the evaluation page in the app shows. Each panel has its own endpoint. Roboflow runs an evaluation when a training finishes, and you can run one again from the app. Every endpoint needs an API key with the model-eval:read scope.

Start with List Model Evaluations to find an evalId, then poll Get a Model Evaluation until status is done. The panel endpoints return 409 until then. Every panel response repeats evalId, project, versionId, and modelId, so you can link to the evaluation in the app at https://app.roboflow.com/{workspace}/{project}/evaluation/{versionId}.

To compare several models on one version, see Compare Trained Models.

List Model Evaluations

Returns a short record for each evaluation, without metrics. Filter by one of project, version, or model; sending more than one returns 400. To match an evaluation to its model, compare modelId with the model ID you use for inference.

List model evaluations

get

Lists the model evaluations in a workspace, newest first. Filter by one of project, version, or model per request. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

Query parameters
projectstringOptional

Project URL slug. An unknown slug returns an empty list.

Example: chess-pieces-fmhpz
versionstringOptional

Dataset version number. Alias: versionId.

Example: 4
modelstringOptional

Model ID, with or without the workspace prefix. Alias: modelId. An unknown ID returns an empty list.

Example: my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1
statusstring · enumOptional

Return only evaluations with this status.

Example: donePossible values:
limitinteger · min: 1 · max: 200Optional

Page size. Values above 200 are lowered to 200.

Default: 50
Responses
200

Successful response.

application/json
get/{workspace}/model-evals
GET /{workspace}/model-evals HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evals": [
    {
      "evalId": "huUF720inUcymARwqAGK",
      "project": "chess-pieces-fmhpz",
      "versionId": "4",
      "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
      "status": "done",
      "createdAt": "2026-04-27T20:04:10.904Z",
      "medianLatencyMs": 11.9
    }
  ]
}

Get a Model Evaluation

Returns one evaluation. Once status is done, the response adds summary with the headline metrics: mAP for detection tasks, mIoU for semantic segmentation.

Get a model evaluation

get

Returns one evaluation. When status is done, the response adds summary with the headline test-split metrics. Poll this endpoint until status is done before you call the other evaluation endpoints. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

statusstring · enumOptional

pending while the model compiles for inference, running once the evaluation job starts, then done or failed.

Possible values:
createdAtstring · nullableOptional
get/{workspace}/model-evals/{evaluationId}
GET /{workspace}/model-evals/{evaluationId} HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "status": "done",
  "createdAt": "2026-04-27T20:04:10.904Z",
  "summary": {
    "mAP": 0.924,
    "mIoU": null,
    "precision": 0.85,
    "recall": 0.85,
    "medianLatencyMs": 11.9
  }
}

Get mAP Results

Returns the main metrics for each split, as shown in the metrics per split panel. Read taskType first: it tells you whether the splits hold mAP fields or mIoU and pixel-level fields.

Get mAP results

get

Returns the main accuracy metrics for each split. Check taskType before you read the splits: detection splits have mAP by object size and class, semantic segmentation splits have mIoU and pixel-level metrics by class. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

taskTypestring · enumOptional

Which shape each split has. object-detection-like splits carry mAP fields. semantic-segmentation and action-recognition splits carry mIoU and pixel (or timeline) metrics. anomaly-detection splits carry auroc, averagePrecision, precision, recall, f1, falsePositiveRate, and threshold.

Possible values:
get/{workspace}/model-evals/{evaluationId}/map-results
GET /{workspace}/model-evals/{evaluationId}/map-results HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "taskType": "object-detection-like",
  "splits": {
    "test": {
      "map50": 0.924,
      "map50_95": 0.756,
      "map75": 0.924,
      "byObjectSize": {
        "small": {
          "map50": 0.904,
          "map50_95": 0.648,
          "map75": 0.904
        },
        "medium": {
          "map50": 0.991,
          "map50_95": 0.857,
          "map75": 0.991
        },
        "large": null
      },
      "perClass": {
        "Car-rims": {
          "map50": 0.924,
          "map50_95": 0.756,
          "map75": 0.924,
          "byObjectSize": {
            "small": {
              "map50": 0.9,
              "map50_95": 0.65,
              "map75": 0.85
            },
            "medium": {
              "map50": 0.99,
              "map50_95": 0.85,
              "map75": 0.99
            },
            "large": null
          }
        }
      }
    }
  }
}

Get a Confidence Sweep

Returns the curves behind the production metrics explorer: precision, recall, and F1 at each confidence threshold, and the threshold with the best F1. Use it to choose the confidence threshold to deploy with.

Get a confidence sweep

get

Returns precision, recall, and F1 at each confidence threshold, with the F1-optimal threshold, for each split and class. Use it to pick a confidence threshold for deployment. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

get/{workspace}/model-evals/{evaluationId}/confidence-sweep
GET /{workspace}/model-evals/{evaluationId}/confidence-sweep HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "splits": {
    "test": {
      "perThreshold": {
        "0.20": {
          "precision": 0.45,
          "recall": 0.92,
          "f1": 0.605
        },
        "0.37": {
          "precision": 0.85,
          "recall": 0.85,
          "f1": 0.85
        },
        "0.50": {
          "precision": 0.91,
          "recall": 0.78,
          "f1": 0.84
        }
      },
      "optimalThreshold": 0.37,
      "optimalMetrics": {
        "precision": 0.85,
        "recall": 0.85,
        "f1": 0.85
      },
      "perClass": {
        "Car-rims": {
          "perThreshold": {
            "0.37": {
              "precision": 0.85,
              "recall": 0.85,
              "f1": 0.85
            }
          },
          "optimalThreshold": 0.37,
          "optimalMetrics": {
            "precision": 0.85,
            "recall": 0.85,
            "f1": 0.85
          }
        }
      }
    }
  }
}

Get Performance by Class

Returns the performance by class panel for one split. Each class reports precision, recall, and F1 at its own best threshold.

Get performance by class

get

Returns metrics for each class on one split. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Query parameters
splitstring · enumOptional

Split to report. all is not accepted, because per-class metrics use a per-split threshold.

Default: testPossible values:
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

taskTypestringOptional

object-detection-like, semantic-segmentation, or action-recognition. Says which fields each class has.

splitstringOptional
iouThresholdnumber · nullableOptional

IoU threshold used, when the evaluation reports one.

get/{workspace}/model-evals/{evaluationId}/performance-by-class
GET /{workspace}/model-evals/{evaluationId}/performance-by-class HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "taskType": "object-detection-like",
  "split": "test",
  "iouThreshold": null,
  "classes": [
    {
      "className": "Car-rims",
      "map50": 0.924,
      "map50_95": 0.756,
      "map75": 0.924,
      "precision": 0.85,
      "recall": 0.85,
      "f1": 0.85,
      "optimalThreshold": 0.37
    },
    {
      "className": "music-note",
      "map50": null,
      "map50_95": null,
      "map75": null,
      "precision": 0,
      "recall": 0,
      "f1": 0,
      "optimalThreshold": 0.5
    }
  ]
}

Get a Confusion Matrix

Returns the confusion matrix panel. Each confidence threshold gives a different matrix, so pass confidence to match the threshold you deploy with.

Get a confusion matrix

get

Returns the confusion matrix, built from the per-image results at one confidence threshold. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Query parameters
splitstring · enumOptional

Split to count. all adds up every split.

Default: allPossible values:
confidenceinteger · max: 100Optional

Confidence threshold as an integer percentage from 0 to 100. Selects which saved report to read. Leave it out for the default report (usually 20). A threshold the evaluation did not save returns 400 invalid_confidence with the saved thresholds in available.

Example: 37
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

splitstringOptional

Split the matrix covers. all adds up train, valid, and test.

confidenceThresholdnumber · nullableOptional

Confidence threshold of the report, from 0 to 1.

classesstring[]Optional

Class names, in matrix order. The last entry is usually background.

get/{workspace}/model-evals/{evaluationId}/confusion-matrix
GET /{workspace}/model-evals/{evaluationId}/confusion-matrix HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "split": "test",
  "confidenceThreshold": 0.2,
  "classes": [
    "Car-rims",
    "music-note",
    "background"
  ],
  "matrix": [
    [
      20,
      0,
      0
    ],
    [
      0,
      0,
      0
    ],
    [
      80,
      0,
      0
    ]
  ]
}

Get Vector Analysis

Returns the vector analysis panel: images grouped by visual similarity, with the model's average F1, precision, and recall on each group. Look for clusters with a low f1Mean to find kinds of images where the model struggles. To see which images are in a cluster, use Get Image Predictions.

Get vector analysis

get

Returns the clusters of visually similar images and the model's average metrics on each, so you can find groups of images where the model struggles. Each image's cluster and coordinate come from Get Image Predictions. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Query parameters
confidenceinteger · max: 100Optional

Confidence threshold as an integer percentage from 0 to 100. Selects which saved report to read. Leave it out for the default report (usually 20). A threshold the evaluation did not save returns 400 invalid_confidence with the saved thresholds in available.

Example: 37
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

confidenceThresholdnumber · nullableOptional

Confidence threshold of the report, from 0 to 1.

get/{workspace}/model-evals/{evaluationId}/vector-analysis
GET /{workspace}/model-evals/{evaluationId}/vector-analysis HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "confidenceThreshold": 0.2,
  "clustering": {
    "method": "hdbscan",
    "nClusters": 54,
    "metrics": {
      "noiseRatio": 0.078,
      "silhouetteScore": 0.489
    },
    "parameters": {
      "min_cluster_size": 2,
      "min_samples": 1,
      "cluster_selection_method": "eom",
      "metric": "euclidean"
    },
    "processingTimeSeconds": 8.36
  },
  "preprocessing": {
    "method": "umap",
    "originalDimensions": 768,
    "targetDimensions": 10,
    "nNeighbors": 30,
    "minDistance": 0.05
  },
  "clusters": [
    {
      "id": -1,
      "numImages": 15,
      "splitDistribution": {
        "train": 12,
        "valid": 2,
        "test": 1
      },
      "metrics": {
        "f1Mean": 0.462,
        "f1Std": 0.219,
        "f1Min": 0.129,
        "f1Max": 0.8,
        "precisionMean": 0.33,
        "recallMean": 0.952,
        "perplexityMean": null,
        "perplexityStd": null,
        "perplexityMin": null,
        "perplexityMax": null
      },
      "sampleImages": [
        "img1.jpg",
        "img2.jpg"
      ]
    }
  ]
}

Get Image Predictions

Returns one record per evaluated image, with its counts, metrics, cluster, and confusion entries. Results are paged with limit and offset.

To inspect one image, pass imageName or evalInputId with includeGeometry=true. The record then includes the saved predictions and ground truth, so you can tell whether a failure is a model error or a label error.

Get image predictions

get

Returns per-image results: true and false positive counts, precision, recall, F1, cluster, and confusion entries. Each page reads the whole saved report, so prefer a few large pages over many small ones. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Query parameters
splitstring · enumOptional

Split to return.

Default: allPossible values:
confidenceinteger · max: 100Optional

Confidence threshold as an integer percentage from 0 to 100. Selects which saved report to read. Leave it out for the default report (usually 20). A threshold the evaluation did not save returns 400 invalid_confidence with the saved thresholds in available.

Example: 37
imageNamestring · min: 1 · max: 1024Optional

Return only the records for this image name.

Example: abc.jpg
evalInputIdstring · min: 1 · max: 256Optional

Return only the record for this evaluated input. Use it when one image has several inputs (ex: tiles or crops).

includeGeometrybooleanOptional

Send true to add geometry to the record: the saved predictions, ground truth, class names, and a temporary image URL. Needs imageName or evalInputId, and the filters must match one record.

Default: false
limitinteger · min: 1 · max: 1000Optional

Page size. Values above 1000 are lowered to 1000.

Default: 200
offsetintegerOptional

Records to skip before the page starts.

Default: 0
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

splitstringOptional
confidenceThresholdnumber · nullableOptional

Confidence threshold of the report, from 0 to 1.

totalImagesintegerOptional

Records that match the filters, before paging.

offsetintegerOptional
limitintegerOptional
get/{workspace}/model-evals/{evaluationId}/image-predictions
GET /{workspace}/model-evals/{evaluationId}/image-predictions HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "split": "test",
  "confidenceThreshold": 0.2,
  "totalImages": 192,
  "offset": 0,
  "limit": 50,
  "images": [
    {
      "imageId": "1QKLCUsfAzFiCIb6YCJj",
      "imageName": "abc.jpg",
      "evalInputId": null,
      "split": "test",
      "augmentations": 2,
      "cluster": {
        "id": 4,
        "embedding2D": [
          7.49,
          -5.14
        ]
      },
      "stats": {
        "truePositives": 2,
        "falsePositives": 7,
        "falseNegatives": 0,
        "precision": 0.222,
        "recall": 1,
        "f1": 0.364
      },
      "confusion": [
        [
          0,
          0,
          2
        ],
        [
          2,
          0,
          7
        ]
      ]
    }
  ]
}

Get Recommendations

Returns the model improvement recommendations panel. Roboflow generates recommendations when a training finishes, and this endpoint only reads them. A finished evaluation without recommendations returns 200 with generated: false, not an error.

Get evaluation recommendations

get

Returns the model improvement recommendations for a finished evaluation. It does not generate them. When none exist yet, the response is 200 with generated set to false. Requires model-eval:read.

Authorizations
AuthorizationstringRequired

Roboflow API key or OAuth access token.

Path parameters
workspacestringRequired

Workspace URL slug.

evaluationIdstringRequired

Model evaluation ID.

Example: huUF720inUcymARwqAGK
Responses
200

Successful response.

application/json
evalIdstringOptional

Model evaluation ID.

projectstring · nullableOptional

Project URL slug. null if the project was deleted.

versionIdstring · nullableOptional

Dataset version number.

modelIdstring · nullableOptional

Model ID ({workspace}/{model-slug}), the same ID the models and inference endpoints use. {project}/{version} for an evaluation that ran on a version instead of one model. null if the model can no longer be found.

generatedbooleanOptional

False when the evaluation has no recommendations yet. The other fields are then left out.

generatedAtstring · nullableOptional

When the recommendations were generated.

get/{workspace}/model-evals/{evaluationId}/recommendations
GET /{workspace}/model-evals/{evaluationId}/recommendations HTTP/1.1
Host: api.roboflow.com
Authorization: Bearer YOUR_SECRET_TOKEN
Accept: */*
{
  "evalId": "huUF720inUcymARwqAGK",
  "project": "chess-pieces-fmhpz",
  "versionId": "4",
  "modelId": "my-workspace/chess-pieces-fmhpz-4-rfdetr-small-t1",
  "generated": true,
  "generatedAt": "2026-04-27T20:05:37.512Z",
  "recommendations": {
    "summary": {
      "confidenceThreshold": 37,
      "split": "test",
      "generatedAt": "2026-04-27T20:05:37.512Z",
      "count": 2,
      "f1": 0.85,
      "precision": 0.85,
      "recall": 0.85
    },
    "items": [
      {
        "id": "56bcd423-38ff-45f9-b3e0-662a71ce44e6",
        "type": "missed_detection",
        "analysis": {
          "affected_class": "Car-rims",
          "count": 3
        }
      },
      {
        "id": "150e49a8-3a61-479a-9e18-3eb751494a70",
        "type": "class_imbalance",
        "analysis": {
          "affected_class": "Car-rims",
          "current_count": 20,
          "total_gt_instances": 20,
          "median_count": 10
        }
      }
    ]
  }
}

MCP Server

Connect your AI agent to the MCP Server and it can review how a model performed with these tools:

Tool
Description

model_evals_list

List model evaluations in the workspace.

model_evals_get

Get the top-level summary for one evaluation.

model_evals_get_map_results

Get per-split mAP results.

model_evals_get_confusion_matrix

Get the confusion matrix.

model_evals_get_performance_by_class

Get per-class performance metrics for one split.

model_evals_get_recommendations

Get generated recommendations for the evaluation, if available.

Last updated

Was this helpful?