Flux RouterDocumentation

Models

Where to find the live Flux model catalogue, the ids you can send, and how to choose between flux-auto, a lane alias and a pinned model.

One Flux key and one endpoint reach every model in the Flux catalogue. The catalogue changes as models are added and retired, so this page does not list them. It tells you where the live list is and which ids to send.

See the live catalogue

  • In the browser: Models shows every model grouped by lab, with the id to pin it, read from the live catalogue.
  • From the API: GET /v1/models returns every id you can put in the model field. Treat it as the source of truth.
bash
curl https://api.fluxrouter.ai/v1/models \
  -H "Authorization: Bearer $FLUX_API_KEY"

Three ways to choose a model

What you sendWhat happensWhat you pay
flux-autoFlux picks among vetted models for each kind of task, cost-awareThe Standard lane rate, or the Fast lane rate when a Fast model answers. Never more than Standard
flux-fast, flux-standard, flux-reasoningFlux picks a model within that laneThat lane's rate
a flux-pinned-* idThe same backing model every timeThat model's provider rate plus the Flux margin

Rates are on Pricing. If you do not know which to pick, use flux-auto. Lanes explains the three lanes, and Let Flux pick, or pin a model helps you decide.

bash
# Let Flux pick
curl https://api.fluxrouter.ai/v1/chat/completions \
  -H "Authorization: Bearer $FLUX_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "flux-auto",
    "messages": [{ "role": "user", "content": "Summarise this PR." }]
  }'

To pin a model, copy its flux-pinned-* id from Models or from GET /v1/models and send it as model.

Abilities have their own ids

Images, transcription and the other abilities use their own endpoints and ids, not the text lanes. See Give your agent abilities, Make and edit images and Transcribe audio.

See which model answered

Responses name the model that served them in the X-Flux-Model header. See Receipts.