Batch Jobs

Upload JSONL files, create asynchronous Chat Completions or Responses batches, and download results. OpenAI, Claude, and Gemini batches cost 50% less than normal API usage; open-source models cost the same as normal API usage or less. Each model page shows its batch price.

Batch uploads use https://api.nano-gpt.com/api/v1. OpenAI, Claude, and Gemini token usage is billed at half the normal synchronous API rate; open-source model rows run through normal routing at the normal price and accept provider, tools, and response_format. Keep every row on the same model and endpoint. Chat rows require max_tokens or max_completion_tokens; Responses rows require max_output_tokens and currently support direct OpenAI models without provider-hosted tools. Embeddings rows take input as a string or an array of strings and currently support text-embedding-3-small, text-embedding-3-large, and text-embedding-ada-002.

Recent Jobs

Active

Completed

API key

Base URL

https://api.nano-gpt.com/api/v1

JSONL batch file

Batch endpoint

Every file row must use /v1/chat/completions, matching the selected endpoint.

No batch jobs yet

Upload a JSONL file above to create the first batch for this API key.

Need the raw API flow? Use the docs with https://api.nano-gpt.com/api/v1.