cjg007

I've been using v4-flash without vision for this months — it's my go-to for code tasks. Now with vision, I'm wondering: if this model can do everything the text-only version does (plus see images), why keep the text-only one around?

Is it just cost/latency? Or is there something text-only does better?

It seems most people directly issue review commands to the LLM, whereas my approach is to use different skills to perform repeated cross-checks, and after each check I add corresponding tests to prevent regressions.

DeepSeek raised their prices. Oh my god.

Besides, projects with lots of dependencies are taking on more risk than they realize. If one dependency gets compromised, you have no idea how many projects are affected until it's too late.

Before DeepSeek-V4-Pro-0813's price goes up, I expect a surge of frantic traffic — hope the servers can hold up.