Skip to content

MiMo V2 Flash VRAM requirements

How much VRAM MiMo V2 Flash needs
4-bit weights: 175 GBWith 8K context: 176 GBWith 32K context: 176 GB
Parameters
310B
Active per token
16.4B
Longest context
256K
Released
Dec 2025
License
MIT
Strongest at
Not rated

How much VRAM MiMo V2 Flash needs

Memory MiMo V2 Flash needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
As released (FP8)291 GB293 GB293 GB295 GB
6-bit237 GB238 GB238 GB240 GB
5-bit205 GB206 GB207 GB209 GB
4-bit175 GB176 GB176 GB179 GB
3-bit141 GB142 GB143 GB145 GB
2-bit108 GB109 GB110 GB112 GB
1.6-bit59.5 GB60.6 GB61.1 GB63.2 GB

One conversation, 16-bit context cache, runtime overhead included.

GPUs and Macs that run MiMo V2 Flash

The best way MiMo V2 Flash runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 4060Too bigPDF: MiMo V2 Flash on RTX 4060 report
RTX 5060Too bigPDF: MiMo V2 Flash on RTX 5060 report
RTX 3070Too bigPDF: MiMo V2 Flash on RTX 3070 report
RTX 3080 10 GBToo bigPDF: MiMo V2 Flash on RTX 3080 10 GB report
RTX 3060 12 GBToo bigPDF: MiMo V2 Flash on RTX 3060 12 GB report
Arc B580Too bigPDF: MiMo V2 Flash on Arc B580 report
RTX 4070Too bigPDF: MiMo V2 Flash on RTX 4070 report
RTX 4070 SuperToo bigPDF: MiMo V2 Flash on RTX 4070 Super report
RTX 5070Too bigPDF: MiMo V2 Flash on RTX 5070 report
RTX 4060 Ti 16 GBToo bigPDF: MiMo V2 Flash on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GBToo bigPDF: MiMo V2 Flash on RTX 5060 Ti 16 GB report
Arc A770 16 GBToo bigPDF: MiMo V2 Flash on Arc A770 16 GB report
Radeon RX 9070 XTToo bigPDF: MiMo V2 Flash on Radeon RX 9070 XT report
RTX 4070 Ti SuperToo bigPDF: MiMo V2 Flash on RTX 4070 Ti Super report
RTX 4080 SuperToo bigPDF: MiMo V2 Flash on RTX 4080 Super report
RTX 5070 TiToo bigPDF: MiMo V2 Flash on RTX 5070 Ti report
RTX 5080Too bigPDF: MiMo V2 Flash on RTX 5080 report
Radeon RX 7900 XTXToo bigPDF: MiMo V2 Flash on Radeon RX 7900 XTX report
RTX 3090Too bigPDF: MiMo V2 Flash on RTX 3090 report
RTX 4090Too bigPDF: MiMo V2 Flash on RTX 4090 report
Radeon AI PRO R9700Too bigPDF: MiMo V2 Flash on Radeon AI PRO R9700 report
RTX 5090Too bigPDF: MiMo V2 Flash on RTX 5090 report
2× RTX 3090Too bigPDF: MiMo V2 Flash on 2× RTX 3090 report
2× RTX 4090Too bigPDF: MiMo V2 Flash on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GBToo bigPDF: MiMo V2 Flash on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 AdaToo bigPDF: MiMo V2 Flash on RTX 6000 Ada report
L40SToo bigPDF: MiMo V2 Flash on L40S report
2× RTX 5090Too bigPDF: MiMo V2 Flash on 2× RTX 5090 report
4× RTX 3090Too bigPDF: MiMo V2 Flash on 4× RTX 3090 report
RTX PRO 6000 BlackwellToo bigPDF: MiMo V2 Flash on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 16 GB report
Mac mini M4 24 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 24 GB report
Mac mini M4 32 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GBToo bigPDF: MiMo V2 Flash on Mac mini M4 Pro 64 GB report
M4 Max 128 GBToo bigPDF: MiMo V2 Flash on M4 Max 128 GB report
M5 Max 128 GBToo bigPDF: MiMo V2 Flash on M5 Max 128 GB report
M3 Ultra 96 GBToo bigPDF: MiMo V2 Flash on M3 Ultra 96 GB report
M3 Ultra 256 GB4-bit256K39 tok/s9.8 sPDF: MiMo V2 Flash on M3 Ultra 256 GB report
M3 Ultra 512 GBFP8256K26 tok/s9.8 sPDF: MiMo V2 Flash on M3 Ultra 512 GB report
M5 Ultra 512 GBFP8256K36 tok/s3.1 sPDF: MiMo V2 Flash on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GBToo bigPDF: MiMo V2 Flash on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GBToo bigPDF: MiMo V2 Flash on Ryzen AI Max+ 395 128 GB report
RTX Spark2-bit240K31 tok/s3.5 sPDF: MiMo V2 Flash on RTX Spark report
DGX Spark2-bit240K28 tok/s3.5 sPDF: MiMo V2 Flash on DGX Spark report
Dual DGX Spark5-bit256K16 tok/s2.6 sPDF: MiMo V2 Flash on Dual DGX Spark report
DGX Station GB3006-bit88K179 tok/s0.3 sPDF: MiMo V2 Flash on DGX Station GB300 report
2× RTX PRO 6000 Blackwell4-bit256K92 tok/s0.5 sPDF: MiMo V2 Flash on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 BlackwellFP8256K62 tok/s0.3 sPDF: MiMo V2 Flash on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 BlackwellFP8256K62 tok/s0.2 sPDF: MiMo V2 Flash on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXMToo bigPDF: MiMo V2 Flash on A100 80 GB SXM report
H100 SXMToo bigPDF: MiMo V2 Flash on H100 SXM report
4× H200 NVLFP8256K128 tok/s0.3 sPDF: MiMo V2 Flash on 4× H200 NVL report
8× H200 NVLFP8256K128 tok/s0.1 sPDF: MiMo V2 Flash on 8× H200 NVL report

Best fit is the highest quality, up to 8-bit, that leaves 16K of context. First reply is the wait for a 2K prompt. Speeds are rough, plus or minus 25%.

How to run it

llama.cpp
llama-server -hf unsloth/MiMo-V2-Flash-GGUF
vLLM
vllm serve XiaomiMiMo/MiMo-V2-Flash
MLX (Mac)
mlx_lm.chat --model mlx-community/MiMo-V2-Flash-4bit

Questions

How much VRAM does MiMo V2 Flash need?

MiMo V2 Flash needs 175 GB for its weights at 4-bit, 176 GB with an 8K context and 176 GB with 32K.

What is the minimum VRAM for MiMo V2 Flash?

About 109 GB at 2-bit with an 8K context, the lowest precision we recommend for a model this size. The smallest machine here that runs it without spilling is the M3 Ultra 256 GB.

How big is MiMo V2 Flash?

310B parameters (16.4B active per token): 291 GB of weights as released (FP8), 175 GB at 4-bit.

Can MiMo V2 Flash run on an RTX 4090?

Not on one card. The smallest machine here that runs it is the M3 Ultra 256 GB.

More models