DeepSeek-V4.1-Flash released; API model id is now deepseek-flash

▼ WHAT HAPPENED

DeepSeek released DeepSeek-V4.1-Flash on September 10, 2026, describing it as the smallest model in a new architecture family with native multimodal visual understanding. The weights are on Hugging Face under the MIT license; the model card describes a multimodal Mixture-of-Experts model with 552B backbone parameters and support for contexts of up to one million tokens. On the API side the changelog states that the previous-generation V4 Flash and V4 Flash Vision Exp have been retired, that the names `deepseek-v4-flash` and `deepseek-v4-flash-vision-exp` are temporarily routed to the new model, and that callers should change the model name to `deepseek-flash`.

▼ OPERATOR ANGLE

If you call the DeepSeek API, change `deepseek-v4-flash` to `deepseek-flash` now; the legacy name is only routed temporarily. For local use this is a multi-GPU server model, not a workstation one: at 552B backbone parameters it is out of reach of consumer hardware even when quantized.

SOURCE: https://api-docs.deepseek.com/updates/[VENDOR-PRESS]

[pulse item] · runlocalai.co/pulse/deepseek-v4-1-flash-open-weights-api-rename