Overview
Qwen2.5-VL-7B-Instruct is Qwen's compact instruction-tuned vision-language model. Its dynamic-resolution visual encoder handles images and video alongside text, with strong document parsing, chart analysis, multilingual OCR, spatial grounding, and event localisation. The Fuji deployment serves the checkpoint's 128K context window.
Strengths
- Native image and video understanding
- Strong OCR, chart, and document parsing for its size
- 128K served context on a compact dense model
Use cases
- Extracting and explaining documents, screenshots, and charts
- Visual question answering and spatial grounding
- Video summarisation and event localisation
Good to know
This deployment accepts one image or one video per prompt and emits text only. It does not expose function calling, thinking, audio input, or guaranteed structured output.
API access
One OpenAI-compatible API for every model in the catalog, plus the CLI and the function SDK inside workflows. Sign up free, add a card for $5 in credits, and these requests work:
curl https://gateway.graphn.ai/v1/chat/completions \
-H "Authorization: Bearer $GRAPHN_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "qwen2.5-vl-7b-instruct",
"messages": [{"role": "user", "content": "Hello"}]
}'graphn model chat qwen2.5-vl-7b-instruct -m "What is GraphN?"
@function
async def describe(image_url: str) -> str:
return await vision.analyze(image_url, "What is in this image?", model="qwen2.5-vl-7b-instruct")