Save on summarization with batch inference

Summarize tickets, transcripts, reviews, and threads the same way: one line per record, one job, and results within 24 hours at pricing below OpenRouter batch inference.

Request access Read the docs

API example

One line per record

Submit the job with HTTP or the OpenAI SDK. Each summary is a standard chat request on a batch line, attributed to the user who owns it.

See chat completions

curl https://us.api.lararouter.com/v1/batches \
  -H "Authorization: Bearer $LARAROUTER_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "input_file_id": "file_9Hs4TnQ2",
    "endpoint": "/v1/chat/completions",
    "completion_window": "24h"
  }'

How it works

One file in, one file out

Everything runs through the same jobs API: upload a JSONL file, create a batch, and download the results. There is no interactive tier with a premium price.

Built for large backlogs

A job has the same shape whether it holds ten records or ten million. Submit the full set and download the output file when it is ready.

Every summary is auditable

The audit log records each line: who submitted it, which model responded, what came back, and what it cost.

Right-sized models, lower cost

Summarization rarely needs a frontier model. Choose the least expensive model that meets your quality bar, then run it at the batch rate.

Models

Models for summarization

Recommended models for summarization, with the batch rate each one bills for input and output.

Savings are measured against OpenRouter batch inference for the same model.

Start summarizing at scale

Upload a file, create a batch, and collect every summary within 24 hours. Enterprise plans deliver in as little as 4 hours.

Request access Compare model pricing

		BatchPer 1M tokens
	Recommended for
Command R 35B	Summarization Document Processing	0,25 $US 23% below OpenRouter batch	0,75 $US 21% below OpenRouter batch
DeepSeek-R1-Distill-Qwen 7B	Summarization	0,10 $US 20% below OpenRouter batch	0,10 $US 20% below OpenRouter batch
DeepSeek-R1-Distill-Qwen 14B	Summarization Document Processing	0,15 $US 21% below OpenRouter batch	0,15 $US 21% below OpenRouter batch
DeepSeek-R1-Distill-Llama 70B	Summarization Document Processing	0,40 $US 20% below OpenRouter batch	0,40 $US 20% below OpenRouter batch
Llama 3.1 8B	Summarization	0,09 $US 18% below OpenRouter batch	0,09 $US 18% below OpenRouter batch
Llama 3.2 8B	Summarization	0,09 $US 18% below OpenRouter batch	0,09 $US 18% below OpenRouter batch
Llama 3.3 70B	Summarization Document Processing	0,44 $US 20% below OpenRouter batch	0,44 $US 20% below OpenRouter batch
Llama 3.1 70B	Summarization Document Processing	0,44 $US 20% below OpenRouter batch	0,44 $US 20% below OpenRouter batch
Mistral 7B	Summarization	0,10 $US 20% below OpenRouter batch	0,10 $US 20% below OpenRouter batch
Ministral 8B	Summarization	0,10 $US 20% below OpenRouter batch	0,10 $US 20% below OpenRouter batch
Mistral Small 3 24B	Summarization Document Processing	0,30 $US 20% below OpenRouter batch	0,30 $US 20% below OpenRouter batch
Mixtral 8×7B	Summarization Document Processing	0,30 $US 20% below OpenRouter batch	0,30 $US 20% below OpenRouter batch
Qwen2.5 7B	Summarization	0,10 $US 20% below OpenRouter batch	0,15 $US 25% below OpenRouter batch
Qwen2.5 14B	Summarization Document Processing	0,10 $US 26% below OpenRouter batch	0,40 $US 20% below OpenRouter batch
Qwen2.5 32B	Summarization Document Processing	0,40 $US 20% below OpenRouter batch	0,40 $US 20% below OpenRouter batch
Qwen2.5 72B	Summarization Document Processing	0,45 $US 22% below OpenRouter batch	0,60 $US 20% below OpenRouter batch