olmOCR-2-7B-1025
Ai2 · allenai/olmOCR-2-7B-1025
Weights license: Apache 2.0 · code and components may use other terms
Quality
No comparable score
We track no public, comparable benchmark that covers this model yet. No score is inferred.
Reported by the publisher
- ArXiv
- 82.9
- Old Scans Math
- 82.1
- Tables
- 84.3
- Old Scans
- 48.3
- Headers and Footers
- 95.7
- Multi column
- 84.3
- Long tiny text
- 81.4
- Base
- 99.7
Values as printed on the card, with the publisher's own settings (card section: olmOCR-Bench Scores); column “olmOCR pipeline v0.4.0 with olmOCR-2-7B-1025”. Not comparable across publishers and not used in any ranking. Model card
Runs on
Fast memory holds it 90
The planning estimate fits the memory the model can use quickly: graphics card VRAM, or unified memory within the macOS GPU share.
Show all 90
Within memory, outside the fast part 26
Within the machine's memory, outside its fast part: on a graphics card the layers that do not fit run on the CPU from PC RAM (the listed RAM, or 32 GB assumed); on a Mac, above the macOS default GPU share. Not a speed figure.
Show all 26
One memory tier short 3
The same machines with the next memory size up would hold it in fast memory.
Hardware profiles grouped by the Q4 planning estimate on each machine (dedicated GPUs reserve 2 GB; unified and system memory 25%, 2 to 8 GB): the fast memory holds it, it is within memory but outside the fast part (layers offloaded to PC RAM, or above the macOS GPU share), or the machine is one memory tier short. Not a measured fit or speed.
Ai2 released olmOCR-2-7B-1025 on 2025-10-21: a vision-language model that answers about images, with 8.3B parameters on the Hugging Face count and the Apache 2.0 license. Languages the publisher declares: en.
License: commercial use allowed. At Q4 the weights load about 6.2 GB before each machine's reserve; image resolution, video frames and context add memory, and the vision encoder needs support in the runtime you pick.
Measured speed
No published measurement for this checkpoint
Tokens per second are shown only when a source publishes them with hardware, build and setup. They are never estimated.
How to run
Runtime compatibility and tokens per second have not been measured for this entry. Published metadata and a memory estimate do not prove a successful local run.
Library in source: transformers
Documented ways to run it
Copied as printed in the linked source (read 2026-10-18); not run by this site.
pip install olmocr>=0.4.0import torch
import base64
import urllib.request
from io import BytesIO
from PIL import Image
from transformers import AutoProcessor, Qwen2_5_VLForConditionalGeneration
from olmocr.data.renderpdf import render_pdf_to_base64png
from olmocr.prompts import build_no_anchoring_v4_yaml_prompt
# Initialize the model
model = Qwen2_5_VLForConditionalGeneration.from_pretrained("allenai/olmOCR-2-7B-1025", torch_dtype=torch.bfloat16).eval()
processor = AutoProcessor.from_pretrained("Qwen/Qwen2.5-VL-7B-Instruct")
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)
# Grab a sample PDF
urllib.request.urlretrieve("https://olmocr.allenai.org/papers/olmocr.pdf", "./paper.pdf")
# Render page 1 to an image
image_base64 = render_pdf_to_base64png("./paper.pdf", 1, target_longest_image_dim=1288)
# Build the full prompt
messages = [
{
"role": "user",
"content": [
{"type": "text", "text": build_no_anchoring_v4_yaml_prompt()},
{"type": "image_url", "image_url": {"url": f"data:image/png;base64,{image_base64}"}},
],
}
]
# Apply the chat template and processor
text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
main_image = Image.open(BytesIO(base64.b64decode(image_base64)))
inputs = processor(
text=[text],
images=[main_image],
padding=True,
return_tensors="pt",
)
inputs = {key: value.to(device) for (key, value) in inputs.items()}
# Generate the output
output = model.generate(
**inputs,
temperature=0.1,
max_new_tokens=50,
num_return_sequences=1,
do_sample=True,
)
# Decode the output
prompt_length = inputs["input_ids"].shape[1]
new_tokens = output[:, prompt_length:]
text_output = processor.tokenizer.batch_decode(
new_tokens, skip_special_tokens=True
)
print(text_output)
# ['---\nprimary_language: en\nis_rotation_valid: True\nrotation_correction: 0\nis_table: False\nis_diagram: False\n---\nolmOCR: Unlocking Trillions of Tokens in PDFs with Vision Language Models\n\nJake Poz']Technical details
What the source and the catalog store for this checkpoint. Missing fields are left out, never zero.
- Checkpoint / source
- allenai/olmOCR-2-7B-1025
- Task
- image-text-to-text
- Architecture
- Qwen2_5_VLForConditionalGeneration
- Library in source
- transformers
- Weight license
- Apache 2.0allenai/olmOCR-2-7B-1025 model card (Hugging Face) · read 2026-10-18source
- Commercial use
- Allowed Apache License 2.0 text (SPDX) · read 2026-10-18license text
- Languages
- enModel card metadata · read 2026-10-25source
- Released
- olmOCR README release history: 'October 21, 2025 - v0.4.0 - New model release' (Ai2, GitHub) · read 2026-10-18source
Every stored field
- Revision checked
- e52d6f090b7a9007afffbbd6ce510876222fea93
- Architecture
- Qwen2_5_VLForConditionalGeneration
- Library
- transformers
- Hub task tag
- image-text-to-text
- Declared base models
- Qwen/Qwen2.5-VL-7B-Instruct
- Formats
- transformers, safetensors
- Access
- Open download
- Publisher kind
- Official
- Repository created
- 2025-10-06creation is not the model's release
- Repository updated
- 2025-10-22a commit is not a release date
- Hub downloads (30 days) · likes
- 76,658 · 159this repository only · read 2026-10-25
- Catalog state
- Hub config or weights found; no inventory benchmark link
Sources
Identity and published specifications come from the checkpoint page. Benchmark scores belong to the named release and do not measure speed on your computer.
- Publisher model card
- License: allenai/olmOCR-2-7B-1025 model card (Hugging Face)
- Release date: olmOCR README release history: 'October 21, 2025 - v0.4.0 - New model release' (Ai2, GitHub)
Curated facts checked: 2026-10-18. The license and release date cite the publisher's sources; parameters, files and task come from the Hugging Face record, and memory is a labelled estimate from its parameter count. A value no source states stays empty; a repository commit is not a model release date.






























































