DeepSeek-V4-Flash-Vision-Exp is now available as an experimental multimodal model, a model that accepts text and images.

Images can be sent as base64, through an external URL, or through the Files API. For billing, each image counts as up to 384 tokens at V4-Flash pricing.

Claim check:

  • DeepSeek-V4-Flash-Vision-Exp is available in the DeepSeek API. (confirmed by the publication itself: evidence; «DeepSeek-V4-Flash-Vision-Exp is now live on the DeepSeek API Platform!»)
  • The model supports mixed text and image input. (confirmed by the publication itself: evidence; «Supports mixed text + image input; images can be provided via base64, external URLs, or the Files API»)
  • Images are billed as up to 384 tokens each at V4-Flash pricing. (confirmed by the publication itself: evidence; «Images are tokenized for billing: up to 384 tokens each, at V4-Flash pricing»)
  • DeepSeek-V4-Flash-Vision-Exp is an experimental multimodal model. (confirmed by the publication itself: evidence; «This experimental multimodal model matches DeepSeek-V4-Flash on text capabilities—including agents, reasoning, and world knowledge.»)
  • Images can be sent as base64, through an external URL, or through the Files API. (confirmed by the publication itself: evidence; «Supports mixed text + image input; images can be provided via base64, external URLs, or the Files API»)

Publications:

Primary sources:

score 26.6 out of 100 · kind: research · update 4