olmOCR-7B-0225-preview

8.29B parameters, images and text in, text out, 32K tokens of context.

Verified 2026-10-25 Rank #543 in LLMs · score 75.1 method v1 Sources 3 Downloads 30d 47.9k · #384 of 619 tracked

Can I run it?5.0GB at INT4Estimatefits a 16 GB cardmethod

Can I use it?Apache 2.0Permissive; patent grantlicence file2026-10-24

Do people use it?47.9kdownloads, 30 days#384 of 619 we track · below the median+0 likes, downloads −41% this week705 likesHugging Face2026-10-24

Worth reading

Picked by rule from 7 tracked pieces: the lab's own post, the most shared, a field report, the latest.

Latest · 12 months ago

OlmOCR 2

olmOCR converts PDFs and image-based documents into clean Markdown using a 7B parameter VLM that requires a GPU.

Local GPU inference requires a recent NVIDIA GPU with at least 12 GB of GPU RAM and 30GB of free disk space.
The v0.2.1 release runs significantly faster because it defaults to FP8 precision and needs much fewer retries per document.

github.com 2025-10-24 5 HN pts · 1 comments 2.1k words about the olmOCR family family

7 best of 7 tracked

The 1 pick above, plus these 6. Top is the radar's score: discussion, source authority and the lab's own posts, fading with age.

ArticleOlmOCR: Open-source tool to extract plain text from PDFsolmocr.allenai.org · 2025-02-25 · 313 HN pts · official
ArticleOlmOCR 2: Unit test rewards for document OCRallenai.org · 2025-10-22 · 3 HN pts · official
ArticleTaming the LLM OCR Wild West: Which Model Works?medium.com · 2025-03-14 · 3 HN pts
ArticleOlmOCR vs. Gemini 2.0 Flash: A Comparison for PDF OCRmedium.com · 2025-03-06 · 2 HN pts
ArticleWhy OlmOCR Changes the OCR Game?freeolmocr.com · 2025-03-05 · 2 HN pts
DiscussionOlmOCR: Open-source PDF to plain text, $190 USD / million pagesnews.ycombinator.com · 2025-02-25 · 1 HN pts

Show all 6

Where it's used

Hugging Face does not say who downloads a model or with what. These are the proxies: API providers listing it, quantized builds, fine-tunes/adapters/merges, and Spaces that load it.

No API provider serves it, so its downloads are people running it themselves: 14 quantized builds and 12 Spaces.

As a service0no API provider listed
On own hardware14quantized builds
Built on77 fine-tunes · 0 adapters · 0 merges
In apps12Hugging Face Spaces

Top Spaces running it: davanstrien/ocr-time-machine ♥ 68 · prithivMLmods/Multimodal-VLM-Thinking ♥ 57

Will it fit?

Weights only, an estimate: leave room for the KV cache. Method

Fits a 16 GB card at INT4

About 5.0 GB at INT4, 9.9 GB at FP8, 20 GB at BF16: weights only, so leave room for the KV cache.

Estimate params × bytes per param × 1.2 · how we estimate

The story so far

Weights, every dated article, and repo commits on one rail.

7 dated articles and 21 repo commits so far

Teal is the weights (2025-01-15), amber triangles are 7 dated articles (each links to it), grey ticks are repo commits.

What people built on it

21 derivatives on the Hub. Which build to run, by runtime: the most downloaded in each class.

21 derivatives on the Hub, most of them quantized

From the Hugging Face model tree; quantized builds in teal, fine-tunes, merges and adapters in greys.

GGUF bartowski/allenai_olmOCR-7B-0225-preview-GGUF for llama.cpp, Ollama, LM Studio 35k downloads

All variants →

Reference

Everything else, folded: the full tables, every source.

What the card says

1 cited sentence

Summary generated, cited

olmOCR-7B-0225-preview is an 8.29B parameter image-text-to-text model fine tuned from Qwen2-VL-7B-Instruct.^[1]^[2]

parameters human
“fine tuned from Qwen2-VL-7B-Instruct”

Generated 2026-10-07 15:12 UTC by parasail-gemma-4-26b-a4b-it; each numeral points at the fact it rests on. Not a verdict.

Use-case evidence

4 use cases

What it is good at 4 use cases

Labels are evidence, not verdicts: measured a third-party leaderboard row for this model; author-reported a number the model author published in the card or eval files; structural a deterministic signal from the name, config or chat template, or a decision-model estimate. Each chip links to its source. All use cases.

Facts

20 fields

Listed in LLMs

Facts

olmOCR-7B-0225-preview has 8.29B parameters. (safetensors metadata, observed 2026-10-24)
Context length is 32,768 tokens (32K). (config.json max_position_embeddings, observed 2026-10-24)
License: Apache 2.0 (friction: none). Permissive; patent grant. (license file, observed 2026-10-24)
Created 2025-01-15, last modified 2025-08-19. (HF API, observed 2026-10-24)

Pipeline tag	image-text-to-text
Library	transformers
Architectures	Qwen2VLForConditionalGeneration
Model type	qwen2_vl
Dtype	bfloat16
Params by dtype	BF16: 8,291,375,616
Size band	M
MoE	no
Context length	32,768 (config.json max_position_embeddings)
Languages	en
License	Apache 2.0 (apache-2.0)
Hub snapshot	2026-10-25

Servability

vLLM · SGLang

Servability

vLLM: supported · Qwen2VLForConditionalGeneration · 7050ef881ff0835f6890a47db1859a24c65481d6SGLang: supported · Qwen2VLForConditionalGeneration · 773cb938fe9484a8f7e7e46fb1b1fe96412c59ed

vLLM registry (7050ef881ff0835f6890a47db1859a24c65481d6) · SGLang registry (773cb938fe9484a8f7e7e46fb1b1fe96412c59ed) · Matched by architecture class name against each project's registry on main; 'supported' means the architecture is registered, not that this checkpoint was tested.

Hardware fit

3 estimates

Hardware fit ESTIMATE

Weights-only memory for olmOCR-7B-0225-preview: params × bytes/param × 1.2 overhead. KV cache and activations are extra. How this is computed.

BF16 (2 bytes/param)	20 GB · Single GPU (<24 GB)
FP8 (1 byte/param)	9.9 GB · Single GPU (<24 GB)
INT4 (0.5 bytes/param)	5.0 GB · Single GPU (<24 GB)

Weights only (params × bytes/param × 1.2 overhead); KV cache and activations extra.

All variants

21 variants

Lineage

Fine-tuned from Qwen/Qwen2-VL-7B-Instruct Family: allenai/olmOCR 21 derivatives on the Hub

Base model: Qwen/Qwen2-VL-7B-Instruct

quantized: 14finetune: 7

Top variants (21 shown, ranked by downloads)

Variant	Relation	Quant	Params	Band	License	Author	Likes	Downloads
bartowski/allenai_olmOCR-7B-0225-preview-GGUF	quantized	GGUF	—	—	apache-2.0	bartowski	7	35k
lmstudio-community/olmOCR-7B-0225-preview-GGUF	quantized	GGUF	—	—	apache-2.0	lmstudio-community	13	26.6k
mradermacher/olmOCR-7B-0225-preview-i1-GGUF	quantized	GGUF	—	—	apache-2.0	mradermacher	0	19.5k
tngtech/olmOCR-7B-faithful	finetune	—	8.29B	M	apache-2.0	tngtech	18	11.8k
mradermacher/olmOCR-7B-0225-preview-GGUF	quantized	GGUF	—	—	apache-2.0	mradermacher	0	4.3k
DevQuasar/allenai.olmOCR-7B-0225-preview-GGUF	quantized	GGUF	—	—	—	DevQuasar	1	4.1k
Adun/olmOCR-7B-thai-v2	finetune	—	8.29B	M	mit	Adun	4	3.8k
Adun/olmOCR-7B-thai-v1	finetune	—	8.29B	M	—	Adun	0	2.8k
Adun/olmOCR-7B-thai-v3.2	finetune	—	8.29B	M	mit	Adun	0	1.6k
Adun/olmOCR-7B-thai-v3	finetune	—	8.29B	M	mit	Adun	0	554
Lucy-in-the-Sky/olmOCR-7B-0225-preview-Q4_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	Lucy-in-the-Sky	1	460
Lucy-in-the-Sky/olmOCR-7B-0225-preview-Q2_K-GGUF	quantized	Q2_K	—	—	apache-2.0	Lucy-in-the-Sky	0	467
Lucy-in-the-Sky/olmOCR-7B-0225-preview-Q8_0-GGUF	quantized	Q8_0	—	—	apache-2.0	Lucy-in-the-Sky	0	369
RCorvalan/olmOCR-7B-0225-preview-Q4_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	RCorvalan	0	277
Lucy-in-the-Sky/olmOCR-7B-0225-preview-Q6_K-GGUF	quantized	Q6_K	—	—	apache-2.0	Lucy-in-the-Sky	0	276
nielsgl/olmOCR-7B-0225-preview-bnb-4bit	quantized	-bnb	7.97B	M	apache-2.0	nielsgl	0	229
sp-embraceable/olmOCR-7B-0225-preview-Q4_K_M-GGUF	quantized	GGUF	—	—	apache-2.0	sp-embraceable	0	171
RCorvalan/olmOCR-7B-0225-preview-Q6_K-GGUF	quantized	Q6_K	—	—	apache-2.0	RCorvalan	0	149
RCorvalan/olmOCR-7B-0225-preview-Q8_0-GGUF	quantized	Q8_0	—	—	apache-2.0	RCorvalan	0	97
karrrr123456/wsrur6utrsrdtryutrdgsdrty	finetune	—	26.8M	S	openrail	karrrr123456	0	0
immaari/Model-Test-08	finetune	—	—	—	—	immaari	0	0

Repository changelog

21 commits

Repository changelog 21 commits

Date	Commit	SHA
2025-08-19	Update README.md	6f102f3
2025-02-25	Update README.md	3123f78
2025-02-21	Update README.md	13c0cb9
2025-02-14	Update README.md	1f2b3f3
2025-02-14	Update README.md	25e9e18
2025-02-14	Update README.md	6993b52
2025-02-14	Update README.md	91b2121
2025-02-14	Update README.md	51f1b47
2025-02-14	Update README.md	4e3f1d7
2025-02-14	Update README.md	dc0bd53
2025-02-14	Update README.md	0c2fd16
2025-02-14	Update README.md	924afd2
2025-02-14	Update README.md	22314c3
2025-02-14	Update README.md	2ad3a7c
2025-01-28	Update config.json	9ecf4b2
2025-01-16	Upload folder using huggingface_hub	bb4a8c0
2025-01-16	Upload Qwen2VLForConditionalGeneration	2ea346c
2025-01-16	Upload Qwen2VLForConditionalGeneration	a5ccec0
2025-01-15	Update README.md	b1d8f7b
2025-01-15	Upload Qwen2VLForConditionalGeneration	d1c854f
2025-01-15	initial commit	e20399f

Articles and coverage

7 pieces

Ranked by source signals (HN, Lobsters, authority domains, official posts) and recency. Summaries are generated from a stored copy of the article text; every sentence cites a span of that text, and nothing is shown that failed the citation check.

Official2

OlmOCR: Open-source tool to extract plain text from PDFs

olmocr.allenai.org2025-02-25officialauthorityHN front pageabout the olmOCR familyHN 313 pts · 40 comments7 words

olmOCR 2 is an open-source tool for converting PDFs to plain text while preserving reading order. The model was fine-tuned on English documents using a multilingual base VLM and trained with synthetic data and unit tests as verifiable rewards. Users can deploy the toolkit on their own GPUs at an estimated cost of just $178 USD per million pages converted, with the demo model available as allenai/olmOCR-2-7B-1025-FP8.written, reviewed, cited

OlmOCR 2: Unit test rewards for document OCR

allenai.org2025-10-22officialauthorityabout the olmOCR familyHN 3 pts · 1 comments876 words

AllenAI releases olmOCR 2, a vision-language model built on Qwen2.5-VL-7B that achieves 82.4 points on olmOCR-Bench for English digitized print documents. The system fine-tunes on a dataset of 270,000 PDF pages and uses Group Relative Policy Optimization with unit test rewards to generate structured Markdown, HTML, and LaTeX in a single pass. An FP8 quantized version delivers 3,400 output tokens/sec on a single H100 GPU, enabling processing of 10,000 pages for less than $2.written, reviewed, cited

The synthetic data pipeline costs $0.12 per page to generate 2,186 PDF pages with 30,381 verifiable test cases.
Training generates 28 completions per document, rewarding only those that pass more unit tests.
The model correctly interprets the date in Abraham Lincoln's January 10, 1864 letter to Major General Hitchcock.

Analysis4

OlmOCR 2

github.com2025-10-24authorityabout the olmOCR familyHN 5 pts · 1 comments2,075 words

olmOCR converts PDFs and image-based documents into clean Markdown using a 7B parameter VLM that requires a GPU. The October 21, 2025 v0.4.0 release boosts olmOCR-bench scores by ~4 points using synthetic data and introduces RL training. Conversion costs less than $200 USD per million pages.written, reviewed, cited

Local GPU inference requires a recent NVIDIA GPU with at least 12 GB of GPU RAM and 30GB of free disk space.
The v0.2.1 release runs significantly faster because it defaults to FP8 precision and needs much fewer retries per document.
Remote inference on Cirrascale costs $0.07 per million input tokens and $0.15 per million output tokens.

Taming the LLM OCR Wild West: Which Model Works?

medium.com2025-03-14about the olmOCR familyHN 3 pts · 1 comments

OlmOCR vs. Gemini 2.0 Flash: A Comparison for PDF OCR

medium.com2025-03-06about the olmOCR familyHN 2 pts · 1 comments

Why OlmOCR Changes the OCR Game?

freeolmocr.com2025-03-05about the olmOCR familyHN 2 pts · 0 comments486 words

Community1

OlmOCR: Open-source PDF to plain text, $190 USD / million pages

news.ycombinator.com2025-02-25about the olmOCR familyHN 1 pts · 0 comments72 words

Similar models

3 models

Similar models

Qwen3-VL-8B-Instruct · 8.77B
same size band (M), same task (image-text-to-text), more providers, permissive license
Qwen2.5-VL-7B-Instruct · 8.29B
same size band (M), same task (image-text-to-text), more providers, permissive license
Qwen3.5-9B · 9.65B
same task (image-text-to-text), more providers, permissive license

Sources

3 sources

Sources

Hub snapshot: 2026-10-25 · HF API: 2026-10-24 21:40 UTC · Providers: 2026-10-24 21:44 UTC · Articles: 2026-10-24 21:46 UTC · Summary: 2026-10-07 15:12 UTC · Last verified 2026-10-25

Follow olmOCR-7B-0225-preview

Watch this model

/feeds/models/allenai--olmOCR-7B-0225-preview.xml(RSS 2.0) · every change to olmOCR-7B-0225-preview
/feeds/families/allenai--olmOCR.xml(RSS 2.0) · every change across the olmOCR family

Email alerts are not open yet; the feeds above are the same data. See feeds for the item schema.