[ batch ]

Batch.

Run asynchronous inference at batch rates on the same USDF balance.

// request

The same request, asynchronous

Submit

Use the :batch model suffix or set batch: true. The worst-case cost is held up front.

Settle

Poll the job by id, then fetch its result: the completion and its receipt. The receipt settles like every other request. Results return within 24 hours.

1. Submit

curl https://api.usdf.fi/v1/chat/completions \
  -H "Authorization: Bearer $USDF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"<model>:batch","messages":[...]}'

2. Job status

curl https://api.usdf.fi/v1/batches/<id> \
  -H "Authorization: Bearer $USDF_API_KEY"

3. Completion and receipt

curl https://api.usdf.fi/v1/batches/<id>/result \
  -H "Authorization: Bearer $USDF_API_KEY"

// rates

Batch rates