inference-history

Get Inference Detail

Changed on

Get a single inference record by ID.

get/v1/inferences/{inference_id}

Request

Path parameters

inference_idstring required

Response

Successful Response

base_modelstring nullable

HuggingFace base model ID

cache_read_tokensinteger nullable

Input tokens served from the provider prompt cache (cache hit).

cache_write_tokensinteger nullable

Input tokens written into the provider prompt cache (cache creation). Zero for providers that bill writes as plain input.

created_atstring date-time required

When the inference was made

error_messagestring nullable

Error detail when status is failed

error_typestring nullable

Failure category when status is failed (validation, timeout, model_not_ready, model_not_found, model_not_supported, capacity_exhausted, internal)

human_corrected_outputunknown
human_feedback_atstring date-time nullable

When human feedback was submitted

human_feedback_bystring nullable

User ID of the member who submitted the current human feedback

human_feedback_notesstring nullable

Optional reviewer notes

human_verdictstring nullable

Human reviewer verdict (correct/incorrect)

idstring required

Unique inference ID

inputstring required

Input text

input_tokensinteger nullable

Non-cached input/prompt tokens, sourced from the metered requests row. None when no billing row was recorded.

latency_msinteger nullable

End-to-end latency in milliseconds

llmaj_judged_atstring date-time nullable

Timestamp when LLMAJ judgment was recorded; None until judged.

llmaj_reasoningstring nullable

LLMAJ judge reasoning/explanation; None until judged.

llmaj_scorenumber nullable

Historical LLMAJ confidence score in [0.0, 1.0].

llmaj_verdict'pass' | 'fail' | 'uncertain' nullable

Historical LLMAJ verdict ('pass', 'fail', or 'uncertain').

metadataobject nullable

Extensible metadata (e.g. LLM judge results)

model_idstring required

Model ID used for inference

model_namestring nullable

Human-readable model name

outputunknown
output_tokensinteger nullable

Output/completion tokens, sourced from the metered requests row.

project_idstring nullable

Project ID the model belongs to

providerstring nullable

Inference provider (aws, modal, etc.). Rows predating a provider removal carry an 'archived_<provider>' label.

sourcestring

Source of the request (api or ui)

statusstring

Inference status: success or failed

taskstring nullable

Task type (legacy; may be NULL for new inferences)

tokensinteger nullable

Token count

training_job_idstring nullable

Training job UUID that produced the model

ttft_msinteger nullable

Streaming time to first visible output chunk in milliseconds; null for non-streaming calls and streams with no visible payload.

user_idstring required

User who made the inference

Changes

    • info

      added the optional property to the response with the status

    • info

      api operation id get_inference_detail_v1_inferences_inference_id_get removed and replaced with get_inference_detail_v1_inferences__inference_id__get