Skip to main content
Qwen3-VL achieves rigorous semantic comprehension of ultra-long documents, supporting native 256K context length that can be extended to 1M tokens. This enables processing of entire books and extensive documents with full recall.

​
Capability Overview

The long document understanding feature enables you to:
  • Process documents up to 256K tokens natively
  • Extend context to 1M tokens with YaRN
  • Handle entire books and lengthy reports
  • Maintain full recall across long contexts
  • Parse complex document structures
  • Extract information from multi-page documents

​
Example Usage

from transformers import AutoModelForImageTextToText, AutoProcessor

model = AutoModelForImageTextToText.from_pretrained(
    "Qwen/Qwen3-VL-235B-A22B-Instruct", dtype="auto", device_map="auto"
)
processor = AutoProcessor.from_pretrained("Qwen/Qwen3-VL-235B-A22B-Instruct")

# For multi-page documents, provide multiple images
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "image": "page1.jpg"},
            {"type": "image", "image": "page2.jpg"},
            {"type": "image", "image": "page3.jpg"},
            {"type": "text", "text": "Summarize the key points from this document."},
        ],
    }
]

inputs = processor.apply_chat_template(
    messages,
    tokenize=True,
    add_generation_prompt=True,
    return_dict=True,
    return_tensors="pt"
)
inputs = inputs.to(model.device)

generated_ids = model.generate(**inputs, max_new_tokens=1024)
generated_ids_trimmed = [
    out_ids[len(in_ids) :] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
]
output_text = processor.batch_decode(
    generated_ids_trimmed, skip_special_tokens=True, clean_up_tokenization_spaces=False
)
print(output_text)

​
Try it Yourself

Explore the full long document understanding cookbook with interactive examples: Open in Colab View on GitHub