GPUs tracked41Price history4 dayssince Sep 3, 2026Price points (24h)10,431 pointsBiggest 24h dropRTX PRO 4000 Blackwell-7.1%Biggest 24h riseGeForce RTX 4080+6.2%Data updated

How Much VRAM Does Kimi K2.7 Code Need to Run Locally?

Kimi K2.7 Code is a MoE model with 1,000B total parameters and 32B active per token. At Q4 it needs 559GB of VRAM to sit entirely on the GPU; with experts offloaded to system RAM it needs only 9GB of VRAM plus 552GB of RAM. No single card in the database fits it; it needs multiple GPUs or the unified memory of a Mac Studio. In offload mode the cheapest card that runs it is the Tesla V100 16GB, at roughly 4 tokens/s.

Kimi K2.7 Code details

Launch dateJun 12, 2026
Hugging Face repomoonshotai/Kimi-K2.7-Code
Revision
ArchitectureMoE
Ollama tag
Min VRAM at Q4559 GB · Q4_K_M · 8K
Experts offloaded9 GB VRAM + 552 GB RAM
Parameters1,000B
MoE32B
Layers61
Hidden size7,168
KV heads64
Head dim192
Max context256K
VendorMoonshot AI

VRAM needed for Kimi K2.7 Code by quantization and context

MoE: all 1,000B parameters must stay resident in memory, while speed is set by the 32B active per token.

QuantizationWeightsGBTotal at 8KGBTotal at 32KGBTotal at 128KGB
1-bit303.9GGUF236.2237.8244.3
2-bit339.5GGUF373.1374.7381.2
3-bit463.9GGUF480.7482.3488.7
4-bit583.7GGUF558.9560.5567
5-bit674.7est.676.3677.9684.3
8-bit594.5GGUF1,038.11,039.71,046.1
FP161,955.8est.1,957.31,958.91,965.4

Total = weights + KV cache (FP16) + 1GB runtime overhead. Weights marked GGUF are measured file sizes from the repository; the rest are estimated from bytes per parameter. Contexts beyond this model's 256K limit show a dash. Each tier takes whichever file of that bit width the repository has, preferring unsloth dynamic quants, so the exact file name varies by model.

Which GPUs can run Kimi K2.7 Code?

Q4_K_M · 8K context · sorted by eBay used price

No single card in the database fits Kimi K2.7 Code: it needs at least 559GB of VRAM at Q4_K_M with an 8K context.

A multi-GPU build needs at least 559GB of VRAM in total.

None of the unified-memory machines in the database fit it either.

GPUs that run Kimi K2.7 Code with experts offloaded to RAM

Expert weights live in system RAM (needs at least 552 GB); VRAM holds only attention layers, shared experts and the KV cache, at least 9 GB at Q4. Speed is bound by RAM bandwidth (estimated at 70 GB/s) and is far slower than a full-VRAM setup.

DeviceVRAMGBUsed priceRAM neededGBEst. t/s
Tesla V100 16GB16$2875523.9
GeForce RTX 2080 Ti11$3505523.9
Radeon RX 7800 XT16$5565523.8
Radeon RX 7900 XT20$6995523.8
Tesla V100 32GB32$7145523.9
Radeon RX 9070 XT16$7505523.8
Radeon RX 7900 XTX24$1,0005523.9
Instinct MI10032$1,0845523.9
GeForce RTX 408016$1,2005523.9
GeForce RTX 5070 Ti16$1,2005523.9
GeForce RTX 309024$1,6205523.9
GeForce RTX 508016$1,7005523.9
RTX PRO 4000 Blackwell24$2,7405523.9
GeForce RTX 409024$3,2005523.9
Radeon PRO W790048$3,4955523.9
RTX PRO 4500 Blackwell32$4,3005523.9
Instinct MI21064$5,0465524
MacBook Pro M4 Max 128GB128$5,4505523.8
A100 40GB PCIe40$6,0005524
GeForce RTX 509032$6,3855524
MacBook Pro M5 Max 128GB128$7,2505523.8
RTX PRO 6000 Blackwell96$16,9855524
Arc Pro B60245523.6
GeForce RTX 5090 D V2245524
Arc Pro B65325523.7
Arc Pro B70325523.7
GeForce RTX 4080 SUPER 32GB325523.9
GeForce RTX 5090 D325524
Radeon AI PRO R9700325523.8
Radeon PRO W7800325523.7
CMP 170HX 40GB405524
GeForce RTX 4090 48GB485523.9
RTX PRO 5000 Blackwell 48GB485524
CMP 170HX 64GB645524
RTX PRO 5000 Blackwell 72GB725524
RTX PRO 6000D845524
DGX Spark 128GB1285523.6
Mac Studio M4 Max 128GB1285523.8
Mac Studio M5 Max 128GB1285523.8
Ryzen AI Max+ 395 128GB1285523.4
Mac Studio M3 Ultra 512GB5125523.9

Run Kimi K2.7 Code with Ollama, llama.cpp or vLLM

Ollama

The Ollama library has no official tag for it yet.

llama.cpp

llama.cpp pulls the GGUF straight from Hugging Face; -c sets the context length.

llama-server -hf unsloth/Kimi-K2.7-Code-GGUF:Q4_K_M -c 8192
vLLM

vLLM serves the original-precision weights, which needs far more memory than GGUF and usually more than one GPU.

vllm serve moonshotai/Kimi-K2.7-Code

File names follow that week's model-library snapshot; the definitions are in the methodology. Methodology

FAQ

How much VRAM does Kimi K2.7 Code need?

At Q4_K_M with an 8K context Kimi K2.7 Code needs about 559GB of VRAM; 672GB leaves comfortable headroom.

What is the cheapest GPU that runs Kimi K2.7 Code?

No single card in the database fits Kimi K2.7 Code; it needs multiple GPUs or a Mac Studio's unified memory.

Can you run Kimi K2.7 Code with Ollama?

Not yet: the Ollama library has no official tag for it. You can load the GGUF with llama.cpp instead.