Skip to content

MiniCPM V 4.7 35B A3B VRAM requirements

How much VRAM MiniCPM V 4.7 35B A3B needs
4-bit weights: 19.9 GBWith 8K context: 21.1 GBWith 32K context: 21.6 GB8-bit, 8K context: 36.0 GB
Parameters
35.2B
Active per token
3B
Longest context
256K
Released
Oct 2026
License
Apache 2.0
Strongest at
Text and images

How much VRAM MiniCPM V 4.7 35B A3B needs

Memory MiniCPM V 4.7 35B A3B needs at each precision and context length
PrecisionWeightsWith 8KWith 32KWith 128K
16-bit65.6 GB66.8 GB67.3 GB69.1 GB
8-bit34.8 GB36.0 GB36.5 GB38.4 GB
6-bit26.9 GB28.1 GB28.6 GB30.4 GB
5-bit23.3 GB24.5 GB25.0 GB26.9 GB
4-bit19.9 GB21.1 GB21.6 GB23.4 GB
3-bit16.0 GB17.2 GB17.7 GB19.6 GB
2-bit12.3 GB13.5 GB14.0 GB15.8 GB
1.6-bit6.8 GB8.0 GB8.4 GB10.3 GB

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

GPUs and Macs that run MiniCPM V 4.7 35B A3B

The best way MiniCPM V 4.7 35B A3B runs on common GPUs, Macs and AI boxes
MachineBest fitContextSpeedFirst replyReport
RTX 4060Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 4060 report
RTX 5060Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 5060 report
RTX 3070Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 3070 report
RTX 3080 10 GBToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 3080 10 GB report
RTX 3060 12 GBToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 3060 12 GB report
Arc B580Too bigPDF: MiniCPM V 4.7 35B A3B on Arc B580 report
RTX 4070Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 4070 report
RTX 4070 SuperToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 4070 Super report
RTX 5070Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 5070 report
RTX 4060 Ti 16 GBToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 4060 Ti 16 GB report
RTX 5060 Ti 16 GBToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 5060 Ti 16 GB report
Arc A770 16 GBToo bigPDF: MiniCPM V 4.7 35B A3B on Arc A770 16 GB report
Radeon RX 9070 XTToo bigPDF: MiniCPM V 4.7 35B A3B on Radeon RX 9070 XT report
RTX 4070 Ti SuperToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 4070 Ti Super report
RTX 4080 SuperToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 4080 Super report
RTX 5070 TiToo bigPDF: MiniCPM V 4.7 35B A3B on RTX 5070 Ti report
RTX 5080Too bigPDF: MiniCPM V 4.7 35B A3B on RTX 5080 report
Radeon RX 7900 XTX4-bit88K146 tok/s0.7 sPDF: MiniCPM V 4.7 35B A3B on Radeon RX 7900 XTX report
RTX 30904-bit88K190 tok/s0.4 sPDF: MiniCPM V 4.7 35B A3B on RTX 3090 report
RTX 40904-bit88K197 tok/s0.2 sPDF: MiniCPM V 4.7 35B A3B on RTX 4090 report
Radeon AI PRO R97006-bit120K98 tok/s0.4 sPDF: MiniCPM V 4.7 35B A3B on Radeon AI PRO R9700 report
RTX 50906-bit120K227 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on RTX 5090 report
2× RTX 30908-bit256K136 tok/s0.2 sPDF: MiniCPM V 4.7 35B A3B on 2× RTX 3090 report
2× RTX 40908-bit256K143 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 2× RTX 4090 report
RTX PRO 5000 Blackwell 48 GB8-bit256K171 tok/s0.2 sPDF: MiniCPM V 4.7 35B A3B on RTX PRO 5000 Blackwell 48 GB report
RTX 6000 Ada8-bit256K138 tok/s0.2 sPDF: MiniCPM V 4.7 35B A3B on RTX 6000 Ada report
L40S8-bit256K129 tok/s0.3 sPDF: MiniCPM V 4.7 35B A3B on L40S report
2× RTX 50908-bit256K200 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 2× RTX 5090 report
4× RTX 30908-bit256K136 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 4× RTX 3090 report
RTX PRO 6000 Blackwell8-bit256K200 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on RTX PRO 6000 Blackwell report
Mac mini M4 16 GBToo bigPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 16 GB report
Mac mini M4 24 GBToo bigPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 24 GB report
Mac mini M4 32 GB3-bit168K47 tok/s9.8 sPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 32 GB report
Mac mini M4 Pro 24 GBToo bigPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 Pro 24 GB report
Mac mini M4 Pro 48 GB6-bit256K59 tok/s4.9 sPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 Pro 48 GB report
Mac mini M4 Pro 64 GB8-bit256K50 tok/s4.9 sPDF: MiniCPM V 4.7 35B A3B on Mac mini M4 Pro 64 GB report
M4 Max 128 GB8-bit256K74 tok/s2.5 sPDF: MiniCPM V 4.7 35B A3B on M4 Max 128 GB report
M5 Max 128 GB8-bit256K79 tok/s0.7 sPDF: MiniCPM V 4.7 35B A3B on M5 Max 128 GB report
M3 Ultra 96 GB8-bit256K85 tok/s1.4 sPDF: MiniCPM V 4.7 35B A3B on M3 Ultra 96 GB report
M3 Ultra 256 GB8-bit256K85 tok/s1.4 sPDF: MiniCPM V 4.7 35B A3B on M3 Ultra 256 GB report
M3 Ultra 512 GB8-bit256K85 tok/s1.4 sPDF: MiniCPM V 4.7 35B A3B on M3 Ultra 512 GB report
M5 Ultra 512 GB8-bit256K102 tok/s0.4 sPDF: MiniCPM V 4.7 35B A3B on M5 Ultra 512 GB report
Ryzen AI Max+ 395 64 GB8-bit256K40 tok/s1.4 sPDF: MiniCPM V 4.7 35B A3B on Ryzen AI Max+ 395 64 GB report
Ryzen AI Max+ 395 128 GB8-bit256K40 tok/s1.4 sPDF: MiniCPM V 4.7 35B A3B on Ryzen AI Max+ 395 128 GB report
RTX Spark8-bit256K56 tok/s0.7 sPDF: MiniCPM V 4.7 35B A3B on RTX Spark report
DGX Spark8-bit256K52 tok/s0.8 sPDF: MiniCPM V 4.7 35B A3B on DGX Spark report
Dual DGX Spark8-bit256K52 tok/s0.4 sPDF: MiniCPM V 4.7 35B A3B on Dual DGX Spark report
DGX Station GB3008-bit256K327 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on DGX Station GB300 report
2× RTX PRO 6000 Blackwell8-bit256K200 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 2× RTX PRO 6000 Blackwell report
4× RTX PRO 6000 Blackwell8-bit256K200 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 4× RTX PRO 6000 Blackwell report
8× RTX PRO 6000 Blackwell8-bit256K200 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 8× RTX PRO 6000 Blackwell report
A100 80 GB SXM8-bit256K214 tok/s0.3 sPDF: MiniCPM V 4.7 35B A3B on A100 80 GB SXM report
H100 SXM8-bit256K264 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on H100 SXM report
4× H200 NVL8-bit256K297 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B on 4× H200 NVL report
8× H200 NVL8-bit256K297 tok/s0.1 sPDF: MiniCPM V 4.7 35B A3B 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 openbmb/MiniCPM-V-4.7-35B-A3B-gguf
vLLM
vllm serve openbmb/MiniCPM-V-4.7-35B-A3B
MLX (Mac)
mlx_lm.chat --model mlx-community/MiniCPM-V-4.7-35B-A3B-mxfp4

Questions

How much VRAM does MiniCPM V 4.7 35B A3B need?

MiniCPM V 4.7 35B A3B needs 19.9 GB for its weights at 4-bit, 21.1 GB with an 8K context and 21.6 GB with 32K. At 8-bit, 36.0 GB.

What is the minimum VRAM for MiniCPM V 4.7 35B A3B?

About 17.2 GB at 3-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 Radeon RX 7900 XTX.

How big is MiniCPM V 4.7 35B A3B?

35.2B parameters (3B active per token): 65.6 GB of weights at 16-bit, 19.9 GB at 4-bit.

Can MiniCPM V 4.7 35B A3B run on an RTX 4090?

Yes. At 4-bit it holds 88K of context and writes about 197 tokens a second.

More models