Skip to main content

DeepSeek-V4-Flash-Vision-Exp Release: Multimodal API Now Live

DeepSeek-V4-Flash-Vision-Exp is now live on the DeepSeek API Platform! ๐Ÿš€

๐Ÿ”น This experimental multimodal model matches DeepSeek-V4-Flash on text capabilitiesโ€”including agents, reasoning, and world knowledge.

๐Ÿ”น On multimodal agent benchmarks, V4-Flash-Vision-Exp makes a major leap over V4-Flash, bringing multimodal agent performance close to Opus-4.8.

Try it with model='deepseek-v4-flash-vision-exp'. DeepSeek Harness 0.1.1 was released today with out-of-the-box support for the new model.

Multimodality unlocks more agent use cases ๐Ÿ‘€โ€‹

V4-Flash-Vision-Exp works smoothly across agent frameworks, combining visual understanding with a wide range of tools to unlock more practical workflows.

Multimodal API support ๐Ÿ”Œโ€‹

๐Ÿ”น Set model='deepseek-v4-flash-vision-exp'

๐Ÿ”น Images are tokenized for billing: up to 384 tokens each, at V4-Flash pricing

๐Ÿ”น Supports Chat Completions, Messages & Responses

๐Ÿ”น Supports mixed text + image input; images can be provided via base64, external URLs, or the Files API

Docs: API Guides - Vision

Files API is now live ๐Ÿ“โ€‹

๐Ÿ”น Free to use

๐Ÿ”น Upload an image once, then reference it by file_id to save request bandwidth

๐Ÿ”น Reuse the same image across requestsโ€”no need to upload it again

Learn more: API Guides - Files API