DeepSeek V4.1 Flash Lands in Tencent's WorkBuddy and CodeBuddy

DeepSeek V4.1 Flash is now running natively inside Tencent's WorkBuddy and CodeBuddy. Tencent AI confirmed the integration, and it comes with a two-week launch promotion.

For developers in China, that means the model runs straight inside Tencent's AI agent and coding workspaces, with no manual proxy setup in the way.

What Happened

DeepSeek V4.1 Flash is built directly into WorkBuddy and CodeBuddy, natively rather than through a proxy. Tencent AI confirmed the rollout and attached an exclusive two-week launch promotion to it.

The model runs on a new Causal Encoder Decoder architecture, with 8B active input and 16B active output. That delivers 300–400+ tokens/sec plus native vision understanding inside developer IDEs.

Tencent and DeepSeek are cutting inference rates for the first 14 days to drive adoption across developer teams. Separately, DeepSeek is rerouting V4 Pro traffic to V4.1 Flash, and enterprise partners like Tencent are making it their default fast-tier backbone.

Why It Matters

If you build on Tencent's coding tools, the fast tier underneath you has quietly changed. V4 Pro requests now land on V4.1 Flash whether you picked it or not, so anything you tuned around V4 Pro's behavior is worth a second look.

The discount only covers the first 14 days. Teams weighing this on cost should check what the rate looks like once that window closes, because the promotional pricing is not the long-term number. Vision support inside the IDE is the other shift worth testing: screenshots, diagrams and UI mockups can go in as prompts now, instead of being described in words first.