deeprelay benchmarks

Read the public inference benchmark (price, TTFT, throughput, weights check)

Synopsis

Read the public inference benchmark (GET /v1/benchmarks/*). Each run measures every model in the benchmark matrix on deeprelay and on independently operated deployments of the same weights: list price per 1M tokens, time to first token and output tokens per second over fixed streamed trials, and a weights-identity check that compares top-k token distributions at temperature 0 so a quantized or substituted build cannot pass as the same model. Runs execute on a public CI workflow; every report carries the run id and URL that produced it. No API key is required.

See the benchmark guide for the method and the report schema.

Options

  -h, --help   help for benchmarks

Options inherited from parent commands

      --api-base string   API base URL (override with DEEPRELAY_API_BASE env) (default "https://api.deeprelay.ai/v1")
      --debug             Enable debug logging to stderr
      --no-preflight      Skip the plan/credit check before inference requests (override with DEEPRELAY_NO_PREFLIGHT env)
  -o, --output string     Output format: table|json (default table on TTY, json otherwise) (default "table")

SEE ALSO

• deeprelay - deeprelay command-line interface
• deeprelay benchmarks get - Show one benchmark run by CI run id
• deeprelay benchmarks latest - Show the newest benchmark run
• deeprelay benchmarks runs - List recent benchmark runs with their CI run ids