Vercel's AI Gateway gives applications a hosted route to DeepSeek V4.1 Flash's image-and-text input. A request can include a screenshot or chart alongside a question, with the model returning text about that visual material. Developers select deepseek/deepseek-v4.1-flash through the gateway, which also provides usage tracking and request routing. The Vercel changelog entry therefore adds a multimodalA model that works with more than text — reading images, audio, or video, and sometimes generating them too.Full definition → model to the gateway's existing interface, allowing applications to submit visual context windowThe maximum amount of text a model can consider at once — its working memory for the current conversation or task.Full definition → through the same model-access layer.
Terms in this piece · Glossary
multimodal — A model that works with more than text — reading images, audio, or video, and sometimes generating them too.
context window — The maximum amount of text a model can consider at once — its working memory for the current conversation or task.
tool use — A model's ability to call external functions — run code, search the web, edit files — instead of only generating text.
prompt caching — Reusing the model's processed form of a repeated prompt prefix so subsequent calls skip re-reading it, cutting cost and latency substantially.
Key quotes
“Its new architecture processes input and generates output with separate components, reducing the active computation needed for each stage.”
