Deepseek’s multimodal model is fully rolled out

Deepseek 的多模态模型全量了。

Deepseek fully rolled out a multimodal model, available via the web image-recognition mode. The post says it looks like a separate model; it does not disclose name, size, pricing, or API timing.

Why it matters: HKR-H/K/R all pass, but the X post only confirms web image-recognition access; model name, params, price, and API timing are missing. DeepSeek’s multimodal rollout is strong, but the thin sourcing keeps it in 78–84.