VRAM · Local model
Kimi K2 (1T) VRAM requirements
Kimi K2 (1T) has 1026.5 billion parameters. With an 8K context it needs about 643 GB of memory at Q4_K_M, and 2104 GB at full precision.
- Parameters
- 1026.5B
- Layers
- 61
- Max context
- 262,144
- Experts
- 384 (8 active)
Attention: 61 layers with multi-head latent attention (one compressed 576-value vector per token).
Requirements
VRAM by quantisation and context length
| Quantisation | 4K context | 32K context | 128K context | 256K context |
|---|---|---|---|---|
| FP16 / BF16 | 2103 GB | 2106 GB | 2113 GB | 2122 GB |
| FP8 | 1052 GB | 1054 GB | 1061 GB | 1070 GB |
| Q8_0 (8-bit) | 1118 GB | 1120 GB | 1127 GB | 1136 GB |
| Q6_K | 863 GB | 865 GB | 872 GB | 881 GB |
| Q5_K_M | 750 GB | 752 GB | 759 GB | 768 GB |
| Q4_K_M | 643 GB | 645 GB | 652 GB | 662 GB |
| MXFP4 | 559 GB | 561 GB | 568 GB | 578 GB |
| Q3_K_M | 526 GB | 528 GB | 535 GB | 545 GB |
| Q2_K | 416 GB | 418 GB | 425 GB | 434 GB |
Weights + FP16 KV cache + overhead (10%, at least 1 GB), batch 1. See the formula and assumptions.
Hardware
What can run it at Q4_K_M
No single GPU in our list has enough memory at Q4_K_M with an 8K context. No Mac can hold it at this setting with the default GPU memory limit.
It’s a mixture-of-experts model: all 384 experts must be in memory, but only 8 run for each token, so it generates faster than a dense model of the same size.
Calculator
Try other settings
1026.5B parameters, 384 experts (8 used per token) · 262,144-token context · model card
GGUF; the most popular balance of size and quality.
Requests served at the same time.
- Weights
- 584 GB
- KV cache
- 0.5 GB
- Overhead
- 58 GB
- Mixture of experts: all 384 experts must be in memory, but only 8 run for each token, so it generates faster than a dense model of this size.
| GPU | Memory | Runs it? |
|---|---|---|
| GeForce RTX 4060 | 8 GB | No |
| GeForce RTX 3060 | 12 GB | No |
| GeForce RTX 4070 | 12 GB | No |
| GeForce RTX 5070 | 12 GB | No |
| GeForce RTX 4060 Ti 16GB | 16 GB | No |
| GeForce RTX 4080 Super | 16 GB | No |
| GeForce RTX 5060 Ti 16GB | 16 GB | No |
| GeForce RTX 5070 Ti | 16 GB | No |
| GeForce RTX 5080 | 16 GB | No |
| GeForce RTX 3090 | 24 GB | No |
| GeForce RTX 4090 | 24 GB | No |
| Radeon RX 7900 XTX | 24 GB | No |
| GeForce RTX 5090 | 32 GB | No |
| L4 | 24 GB | No |
| L40S | 48 GB | No |
| A100 80GB | 80 GB | No |
| H100 80GB | 80 GB | No |
| RTX PRO 6000 Blackwell | 96 GB | With 7 GPUs |
| H200 | 141 GB | With 5 GPUs |
| Instinct MI300X | 192 GB | With 4 GPUs |
Apple Silicon Macs (unified memory)
- 16 GB: doesn’t fit
- 24 GB: doesn’t fit
- 32 GB: doesn’t fit
- 36 GB: doesn’t fit
- 48 GB: doesn’t fit
- 64 GB: doesn’t fit
- 96 GB: doesn’t fit
- 128 GB: doesn’t fit
- 192 GB: doesn’t fit
- 256 GB: doesn’t fit
- 512 GB: doesn’t fit
Filled: fits as is. Outlined: fits after raising the GPU memory limit (leaving 8 GB for macOS). macOS gives the GPU two-thirds of memory up to 32 GB and three-quarters above by default.
FAQ
Frequently asked questions
How much VRAM does Kimi K2 (1T) need?
With an 8K context, about 643 GB at Q4_K_M, 1118 GB at 8-bit (Q8_0) and 2104 GB at full 16-bit precision. That covers the weights, the KV cache and runtime overhead.
Can Kimi K2 (1T) run on a 24 GB GPU?
Not on a single 24 GB card: even at Q2_K it needs about 416 GB. At Q4_K_M you’d need 27 × 24 GB cards.
Can I run Kimi K2 (1T) on a Mac?
Not at Q4_K_M: it needs about 643 GB, more than the largest Mac can give the GPU by default.
How much memory does Kimi K2 (1T)’s context use?
Its KV cache grows by about 67 MB for every 1,000 tokens of context in FP16. At its full 262,144-token context the cache is about 17 GB. Quantising the cache to 8-bit roughly halves it.
Related