Classify and tag
Label support tickets, reviews, or catalog items by the thousands, with JSON output that matches your schema.
batchrate
Start free with 10M tokensLLM inference for jobs that can wait.
Run batch jobs on Qwen3.6 35B for $0.05 per 1M input tokens and $0.40 per 1M output tokens, with results typically back within hours. It's OpenAI-compatible, so your existing code works once you change the base URL.
10M free tokens after email verification. No card needed. Free tokens apply to chat models; see pricing for details.
What people batch
If the job can wait a few hours, batch it.
Extract structured data
Turn PDFs, Word files, and web pages into clean JSON. Upload the documents and Batchrate splits them into requests for you.
Summarize at scale
Condense reports, research papers, transcripts, and long threads in one run.
Read images and scans
Caption product photos, read receipts and screenshots, and pull fields from scanned PDFs with Qwen3.6 35B.
Transcribe audio
Turn calls, podcasts, and meetings into text with Whisper Large v3 Turbo, then summarize or tag the transcripts in a follow-up batch.
Generate and evaluate
Write product descriptions and translations, build synthetic training data, or run evals across your whole test set.
Works with OpenAI's Files and Batches API. JSON mode and JSON Schema output on both chat models.
How it works
1. Upload a file
Write your requests to a JSONL file in the OpenAI batch format, or drop documents, images, or audio into the console.
2. Start a batch
Create a batch with a 24h completion window. Batchrate queues and runs it, with no rate limits or retry loops to manage. You pay only for requests that complete.
3. Download results
Completed lines come back as an output JSONL file. Failures are written to a separate error file.
Same SDK, different base URL
from openai import OpenAI
client = OpenAI(
api_key="br_live_...",
base_url="https://api.batchrate.ai/v1",
)
batch_file = client.files.create(
file=open("batch.jsonl", "rb"),
purpose="batch",
)
batch = client.batches.create(
input_file_id=batch_file.id,
endpoint="/v1/chat/completions",
completion_window="24h",
)
print(batch.id, batch.status)
# Later: poll every 30 to 60 seconds, or use a webhook.
batch = client.batches.retrieve(batch.id)
if batch.status == "completed":
print(client.files.content(batch.output_file_id).text)
Models: qwen3.6-35b-a3b or qwen3.8-27b. Download a JSONL file you can submit. The request format is in the docs.
Prepaid credits
A verified account receives a one-time credit of 10,000,000 tokens, shared by input and output, and used before paid credits. After that, add credits with Stripe Checkout. New batches are rejected when free tokens plus the available balance cannot cover the estimate.
Estimate a batchPay only for what you run
Prepaid credits, no subscription. Pricing is per token. Jobs run inside a 24h completion window, and you pay only for requests that complete.
See pricingYour files don't stick around
Input files are deleted automatically 7 days after their batch finishes. Results are deleted 30 days after they're created. Want them gone sooner? Delete any file anytime through the API or the console.
Read our data policyGot a job that can wait?
Start with 10M free tokens, no card needed. Change one line of code and run your first batch today.