Hacker News new | ask | show | jobs
by grepfru_it 373 days ago
I am curious the need for 70 t/sec?
2 comments
Aeolun 373 days ago
Waiting minutes for your call to succeed is too frustrating?
link
ekianjo 373 days ago
Depends entirely on the use case. Not every LLM workflow is a chatbot
link
jbellis 373 days ago
no, but if you're not latency sensitive you should probably be using DeepSeek v3 (cheaper than flash, significantly smarter)
link
lostmsu 373 days ago
What makes you believe DeepSeek is smarter than Flash 2.5? It is lower on all leaderboards.
link
jbellis 372 days ago
you're right, I should clarify that I'm talking about no thinking mode, otherwise flash goes from "a bit more expensive than dsv3" to "10x more expensive"
link
cootsnuck 373 days ago
High concurrency voice AI systems.
link
grepfru_it 370 days ago
Why are you self hosting that?
link