Allenai

olmOCR 2 7B

Generally availableOpen weights8K context

Summary

olmOCR 2 7B is an open-weight language model from Allenai, released on 22 Oct 2025. It accepts text and images and generates text, with a 8K-token context window and up to 8K output tokens per response.

Vision-language OCR model from Ai2 (Allen Institute for AI) that extracts text from documents where tables, equations, and poor-quality scans defeat traditional OCR

Last verified 18 Nov 2026. Every figure on this page links to its source.

Specifications

DeveloperAllenai
Release date22 Oct 2025 [models.dev]
StatusGenerally available [models.dev]
WeightsOpen weights [models.dev]
Context window8K (8,192 tokens) [models.dev]
Max output8K tokens [models.dev]
InputText and images [models.dev]
OutputText [models.dev]
Reasoning modeNo [models.dev]
Tool callingNo [models.dev]

olmOCR 2 7B API pricing by provider

USD per million tokens. Sorted by input price.

We do not currently track a paid API listing for this model.

Frequently asked questions

What is the context window of olmOCR 2 7B?

olmOCR 2 7B has a context window of 8,192 tokens (8K) and can generate up to 8,192 tokens in a single response, according to models.dev.

Is olmOCR 2 7B open source?

olmOCR 2 7B is an open-weight model. Check the license terms before commercial use.

When was olmOCR 2 7B released?

Allenai released olmOCR 2 7B on 22 Oct 2025, according to models.dev.

Similar models from other labs

Comparable on output type, weights, release date, price and context window. Prices are the lowest listed, USD per million tokens.

ModelLabReleasedContextInput $/1MOutput $/1MWeights
Llama Nemotron Embed 1B v2NVIDIA16 Oct 20258K––Open
llama-nemotron-embed-vl-1b-v2NVIDIA18 Dec 20258K––Open
Command A TranslateCohere28 Aug 20258K$2.50$10Open
DeepSeek OCR 2DeepSeek27 Jan 20268K$0.03$0.03Open
Tiny Aya EarthCohere17 Feb 20268K––Open
PaddleOCR VLPaddlepaddle16 Oct 202516K$0.02$0.02Open

Change history

  • Added to the directory

Sources