Unverified50% confidenceReleaseExact time
On August 3, DeepSeek open-sourced DeepSeek-V4-Flash-Vision-Exp on HuggingFace, the first experimental multimodal vision model in the V4 series.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/3/2026
First Seen
Valid until: 12/2/2026
Sources
Related Entities
Related Claims
UnverifiedDeepSeek has launched the experimental multimodal model V4-Flash-Vision-EXP on its API platform, with text capabilities on par with V4 Flash.77% similarUnverifiedV4-Flash-Vision-Exp performs roughly on par with V4-Flash in pure-text Agent capability.74% similarUnverifiedV4-Flash-Vision-Exp can read screenshots, parse charts, and complete multimodal Agent tasks in combination with tool calling.72% similarUnverifiedDeepSeek V4 Pro added camera-follow and sound effects features in the 3D parkour game task, performing better than expected.68% similarUnverifiedAccording to tests by Bilibili UP host Jiutian, DeepSeek V4 Flash paired with DSH V0.1.1's native visual capabilities can produce front-end pages and HTML-based PPTs, and generate promotional short videos approaching commercial quality from a single sentence.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/849868API
curl https://kongchang.com/api/v1/knowledge/claims/849868MCP
get_claim(id=849868)