[ NVIDIA ]

Nemotron-3.5-Rerank-8B

Nemotron-3.5-Rerank-8B from NVIDIA: a reranking model that orders documents by how well they answer a query. Every model below its source price, and the cheapest provider wins every request. Pay per token in USDF, with no subscription and no minimum. Requests go to /v1/rerank.

Get an API keyTry in Playground
NVIDIAnemotron-3.5-rerank-8bAvailablererankContext window: 32,768 tokens

Input

$0.05

per 1M tokens

Context window

32,768

tokens

Models with a zero-retention route use it by default. A model marked 'provider may keep data' is served by a provider that may keep request data; choosing it is your choice.

Model ID
nemotron-3.5-rerank-8b
Modality
rerank
Unit
USD per 1M tokens
Endpoints
/v1/rerank

Data policy

Zero retentionNo training

Make your first request

Responses include a receipt with the exact cost. No key yet? Create one.

Without an account, send the same request to /x402/v1/rerank with no credentials and the gateway answers 402 with a quote. Pay per request.

curl https://api.usdf.fi/v1/rerank \
  -H "Authorization: Bearer $USDF_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "nemotron-3.5-rerank-8b", "query": "What is USDF?", "documents": ["USDF is a dollar for compute.", "USDG is the reserve asset."]}'