CC Switch Upgrade: DeepSeek-V4-Flash integrates Codex without the need for backend relay

According to Dongcha Beating monitoring, the open-source model switching tool CC Switch has released v3.19.1, adding support for the DeepSeek-V4-Flash native Responses API. Users can input their DeepSeek API Key to directly invoke this model in Codex.

Previously, CC Switch required running a local proxy on the computer to convert Codex's requests into a format recognizable by DeepSeek. Now requests can be sent directly to the DeepSeek official API, making configuration simpler and eliminating a potential error-prone intermediate layer.

The new version will also automatically inform Codex about the maximum text length this model can handle and which inference modes it supports, so users no longer need to manually supplement the configuration. Currently, only V4-Flash supports direct connection, while V4-Pro has not yet been integrated.