eight cents for a full audit

deepseek v3 flash's new 2x usage on opencode go is real.

i ran a full audit launch. 259,772 tokens. 26% used. $0.08 spent.

that's a simple task report for eight cents.

running this is cheaper than running my own llm on the m3 ultra. it's faster and smarter too.

i still see advantages to local llms. just not for my use case right now.

the m3 ultra with 96gb is almost $6k now. buying it just to run llms, plus the power draw, doesn't feel realistic.