vram.run Models Hardware Providers Cloud State of Inference
API provider data is live · Hardware & cloud pricing curated 2026-02-23

MiMo-V2-Flash

by XiaomiMiMo

310B params · text-generation · 643 likes · 485.7k downloads

MiMo-V2-Flash is a 310B parameter model. At Q4 quantization it requires 155GB of VRAM. It fits on M2 Ultra 192GB (2 tok/s) but performance may be limited.

Inference providers

Provider$/1M in$/1M outThroughput
Novita88 tok/s

GPU compatibility

GPUVRAMQ4 DecodeVerdict
M2 Ultra 192GB192GB2 tok/stight
Install CLI [email protected] Raw data · MIT · API data: live · HW/Cloud data: curated 2026-02-23 · v0.6.0