[FuriouslyAdrift]

We're running Kimi 2.8 on a $107k server and getting around 50k tokens per second or better on most things.

We've already saved money compared to last years token cost on Claude/Gemini

[mdp2021]

> We're running Kimi 2.8 on a $107k server

Equipped with what? Is it CPU based inference, a mix...?

[wongarsu]

Not GP, but my educated guess is that they are running a system with between 4 and 6 MI325 or MI355x or similar AMD GPUs. With the 50k tps as the total figure for all parallel requests. Those cards have a lot of memory for their price, allowing you to push to really high batch sizes while still having a large context size for each request

[FuriouslyAdrift]

4x MI300A in a Gigabyte server

[ajju]

What is your config?

[freediddy]

how many requests per second can the server take?

[FuriouslyAdrift]

That's about it across the whole server. Kimi 2.8 is about as big as we can go. Kimi 3 is massive.

[logicallee]

Are you developing software? Is most of it used on a coding agent? (Like Claude Code or ChatGPT Codex?) If so, what coding agent do you use? If you're not developing software what do you use it for (roughly)?

[FuriouslyAdrift]

Our use cases are data analysis, software development, general AI chat with RAG so far.

[retinaros]

on 1 paralel request maybe

[FuriouslyAdrift]

Correct but even with 30 simultaneous users, we're not seeing any noticable lag compared to AIaaS even with agentic workloads.