Skip to content

MiniMax M2.7 VRAM requirements

How much VRAM MiniMax M2.7 needs
4-bit weights: 129 GBWith 8K context: 132 GBWith 32K context: 138 GB
Parameters
229B
Active per token
10B
Longest context
200K
Released
Apr 2026
License
MiniMax Non-Commercial
Strongest at
Coding, Agents

How much VRAM MiniMax M2.7 needs

Memory MiniMax M2.7 needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
As released (FP8)214 GB217 GB223 GB246 GB
6-bit175 GB178 GB183 GB207 GB
5-bit151 GB154 GB160 GB183 GB
4-bit129 GB132 GB138 GB161 GB
3-bit104 GB107 GB113 GB136 GB
2-bit79.9 GB82.7 GB88.5 GB112 GB
1.6-bit43.9 GB46.8 GB52.6 GB75.9 GB

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

GPUs and Macs that run MiniMax M2.7

The best way MiniMax M2.7 runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 4060Too bigPDF: MiniMax M2.7 on RTX 4060 report
RTX 5060Too bigPDF: MiniMax M2.7 on RTX 5060 report
RTX 3070Too bigPDF: MiniMax M2.7 on RTX 3070 report
RTX 3080 10 GBToo bigPDF: MiniMax M2.7 on RTX 3080 10 GB report
RTX 3060 12 GBToo bigPDF: MiniMax M2.7 on RTX 3060 12 GB report
Arc B580Too bigPDF: MiniMax M2.7 on Arc B580 report
RTX 4070Too bigPDF: MiniMax M2.7 on RTX 4070 report
RTX 4070 SuperToo bigPDF: MiniMax M2.7 on RTX 4070 Super report
RTX 5070Too bigPDF: MiniMax M2.7 on RTX 5070 report
RTX 4060 Ti 16 GBToo bigPDF: MiniMax M2.7 on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GBToo bigPDF: MiniMax M2.7 on RTX 5060 Ti 16 GB report
Arc A770 16 GBToo bigPDF: MiniMax M2.7 on Arc A770 16 GB report
Radeon RX 9070 XTToo bigPDF: MiniMax M2.7 on Radeon RX 9070 XT report
RTX 4070 Ti SuperToo bigPDF: MiniMax M2.7 on RTX 4070 Ti Super report
RTX 4080 SuperToo bigPDF: MiniMax M2.7 on RTX 4080 Super report
RTX 5070 TiToo bigPDF: MiniMax M2.7 on RTX 5070 Ti report
RTX 5080Too bigPDF: MiniMax M2.7 on RTX 5080 report
Radeon RX 7900 XTXToo bigPDF: MiniMax M2.7 on Radeon RX 7900 XTX report
RTX 3090Too bigPDF: MiniMax M2.7 on RTX 3090 report
RTX 4090Too bigPDF: MiniMax M2.7 on RTX 4090 report
Radeon AI PRO R9700Too bigPDF: MiniMax M2.7 on Radeon AI PRO R9700 report
RTX 5090Too bigPDF: MiniMax M2.7 on RTX 5090 report
2× RTX 3090Too bigPDF: MiniMax M2.7 on 2× RTX 3090 report
2× RTX 4090Too bigPDF: MiniMax M2.7 on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GBToo bigPDF: MiniMax M2.7 on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 AdaToo bigPDF: MiniMax M2.7 on RTX 6000 Ada report
L40SToo bigPDF: MiniMax M2.7 on L40S report
2× RTX 5090Too bigPDF: MiniMax M2.7 on 2× RTX 5090 report
4× RTX 30902-bit32K106 tok/s0.6 sPDF: MiniMax M2.7 on 4× RTX 3090 report
RTX PRO 6000 Blackwell2-bit40K149 tok/s0.5 sPDF: MiniMax M2.7 on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 16 GB report
Mac mini M4 24 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 24 GB report
Mac mini M4 32 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GBToo bigPDF: MiniMax M2.7 on Mac mini M4 Pro 64 GB report
M4 Max 128 GB2-bit40K56 tok/s10 sPDF: MiniMax M2.7 on M4 Max 128 GB report
M5 Max 128 GB2-bit40K59 tok/s3.0 sPDF: MiniMax M2.7 on M5 Max 128 GB report
M3 Ultra 96 GBToo bigPDF: MiniMax M2.7 on M3 Ultra 96 GB report
M3 Ultra 256 GB6-bit28K42 tok/s5.9 sPDF: MiniMax M2.7 on M3 Ultra 256 GB report
M3 Ultra 512 GBFP8200K36 tok/s5.9 sPDF: MiniMax M2.7 on M3 Ultra 512 GB report
M5 Ultra 512 GBFP8200K46 tok/s1.8 sPDF: MiniMax M2.7 on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GBToo bigPDF: MiniMax M2.7 on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB2-bit40K33 tok/s6.0 sPDF: MiniMax M2.7 on Ryzen AI Max+ 395 128 GB report
RTX Spark3-bit36K37 tok/s2.1 sPDF: MiniMax M2.7 on RTX Spark report
DGX Spark3-bit36K34 tok/s2.3 sPDF: MiniMax M2.7 on DGX Spark report
Dual DGX SparkFP852K18 tok/s2.8 sPDF: MiniMax M2.7 on Dual DGX Spark report
DGX Station GB300FP896K174 tok/s0.2 sPDF: MiniMax M2.7 on DGX Station GB300 report
2× RTX PRO 6000 Blackwell6-bit24K97 tok/s0.3 sPDF: MiniMax M2.7 on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 BlackwellFP8200K85 tok/s0.2 sPDF: MiniMax M2.7 on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 BlackwellFP8200K85 tok/s0.1 sPDF: MiniMax M2.7 on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXMToo bigPDF: MiniMax M2.7 on A100 80 GB SXM report
H100 SXMToo bigPDF: MiniMax M2.7 on H100 SXM report
4× H200 NVLFP8200K149 tok/s0.2 sPDF: MiniMax M2.7 on 4× H200 NVL report
8× H200 NVLFP8200K149 tok/s0.1 sPDF: MiniMax M2.7 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/MiniMax-M2.7-GGUF
vLLM
vllm serve MiniMaxAI/MiniMax-M2.7
MLX (Mac)
mlx_lm.chat --model mlx-community/MiniMax-M2.7-4bit

Questions

How much VRAM does MiniMax M2.7 need?

MiniMax M2.7 needs 129 GB for its weights at 4-bit, 132 GB with an 8K context and 138 GB with 32K.

What is the minimum VRAM for MiniMax M2.7?

About 82.7 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 4× RTX 3090.

How big is MiniMax M2.7?

229B parameters (10B active per token): 214 GB of weights as released (FP8), 129 GB at 4-bit.

Can MiniMax M2.7 run on an RTX 4090?

Not on one card. The smallest machine here that runs it is the 4× RTX 3090.

More models