Inference

List declared benchmark workloads and engine availability

One entry per declared workload: the engine family, the engine id, the model it would measure, whether the engine can run it now, the provenance its rows would carry (measured, or diagnostic where the evidence cannot qualify) and, when unavailable, the reason. Nothing is loaded or measured by this call.

get/v1/benchmark/workloads

Request

  • Base URL: http://localhost:9998
  • URL: http://localhost:9998/v1/benchmark/workloads
  • Auth: HTTP bearer

Response

The declared workloads with their current capability

workloadstring required
family'llm' | 'image' | 'tts' | 'stt' | 'upscale' | 'embedding' | 'rerank' | 'splat' | 'pose' required
engineId'piper-engine' | 'piper-tts' | 'piper-stt' | 'piper-mlx-tts' | 'piper-mlx-llm' | 'apple-intelligence' | 'apple-vision-ocr' | 'embedding-server' | 'embedding-gemma' | 'coreml-upscale' | 'coreml-pose' | 'image-generation' | 'video-generation' | 'mesh-generation' | 'splat-generation' | 'cloud' | 'harness' required
modelIdstring nullable required
availableboolean required
provenance'measured' | 'estimated' | 'historical' | 'diagnostic' nullable required
reasonstring nullable required

Changes

No recorded changes to this operation across all 1 revision of this API.