sference

The most efficient AI inference stack for async workloads

[sference @sferenceai]

Inkling by @thinkymachines is live on sference from day 0.

975B MoE · 41B active · 256k context

Hosted on European GPUs. Try it via the sference API: https://t.co/n7JXb38fyi

[Jernej Strasner @jernejstrasner]

Who wants access to moonshotai/Kimi-K2.7-Code hosted on European servers by a European company? It's free for now. And you can use it in claude code. Comment "euromaxxing" to get access.