KI

Kimi-K2.5

KimiGGUF by unslothUpdated 2026-01-28
Download on Hugging Face
Parameters1030B
Context256K
Layers61
Architecturedeepseek2
Smallest quant239.5 GB
Quick answer

Kimi-K2.5 needs about 622.8 GB of memory at Q4_K_M (a 621.2 GB file) with an 8K context window, which is more than a 48 GB GPU holds. This repository has 27 quants, from 239.5 GB (TQ1_0) to 2053.2 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 4.3 GB.

Kimi-K2.5: 1 000‑billion‑parameter multimodal agentic model

Kimi-K2.5 is a 1 000‑billion‑parameter model built on the deepseek2 architecture. It supports vision and language tasks and includes agentic capabilities. The model is available in multiple GGUF quantizations.

Good at

  • Vision‑language reasoning
  • Agentic tool use
  • Cross‑modal code generation

Limits

  • No vision support in current GGUF builds
  • Requires large unified memory for small quants
  • License is other, not open source

Best for

  • Visual question answering
  • Multimodal code synthesis
  • Agentic task orchestration

Who it’s for. Developers needing a large multimodal model for research or production. The model is suitable for teams with access to high‑memory GPUs or CPUs.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
TQ1_0239.5 GB (5 parts) 241.1 GB
Too big Get file →
IQ1_S275.9 GB (6 parts) 277.5 GB
Too big Get file →
IQ1_M300.6 GB (7 parts) 302.2 GB
Too big Get file →
IQ2_XXS326.7 GB (7 parts) 328.3 GB
Too big Get file →
IQ2_M345.2 GB (7 parts) 346.8 GB
Too big Get file →
Q2_K373.8 GB (8 parts) 375.4 GB
Too big Get file →
Q2_K_L374.1 GB (8 parts) 375.6 GB
Too big Get file →
Q2_K_XL374.8 GB (8 parts) 376.4 GB
Too big Get file →
IQ3_XXS414.8 GB (9 parts) 416.4 GB
Too big Get file →
Q3_K_S443.2 GB (10 parts) 444.7 GB
Too big Get file →
Q3_K_M489.8 GB (11 parts) 491.4 GB
Too big Get file →
Q3_K_XL490.4 GB (11 parts) 492.0 GB
Too big Get file →
IQ4_XS547.2 GB (12 parts) 548.8 GB
Too big Get file →
IQ4_NL579.1 GB (12 parts) 580.7 GB
Too big Get file →
Q4_0581.2 GB (13 parts) 582.8 GB
Too big Get file →
Q4_K_S583.3 GB (13 parts) 584.9 GB
Too big Get file →
Q4_K_M621.2 GB (13 parts) 622.8 GB
Too big Get file →
Q4_K_XL622.0 GB (13 parts) 623.6 GB
Too big Get file →
Q4_1642.9 GB (13 parts) 644.5 GB
Too big Get file →
Q5_K_S707.0 GB (15 parts) 708.6 GB
Too big Get file →
Q5_K_M728.7 GB (16 parts) 730.3 GB
Too big Get file →
Q5_K_XL731.3 GB (16 parts) 732.9 GB
Too big Get file →
Q6_K842.9 GB (18 parts) 844.5 GB
Too big Get file →
Q6_K_XL877.6 GB (19 parts) 879.2 GB
Too big Get file →
Q8_01091.1 GB (23 parts) 1092.6 GB
Too big Get file →
Q8_K_XL1189.7 GB (25 parts) 1191.3 GB
Too big Get file →
BF162053.2 GB (46 parts) 2054.7 GB
Too big Get file →

For 16 GB VRAM and 256 GB RAM, use the Q4_K_S or Q4_K_M quant, which need roughly 585 GB unified memory. For larger memory budgets, Q6_K or Q8_0 provide higher precision but require >840 GB unified memory. The smallest quant, TQ1_0, needs about 240 GB unified memory.

What GPU do you need for Kimi-K2.5?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: no quant fits at an 8K context.
  • 6 GB VRAM: no quant fits at an 8K context.
  • 8 GB VRAM: no quant fits at an 8K context.
  • 12 GB VRAM: no quant fits at an 8K context.
  • 16 GB VRAM: no quant fits at an 8K context.
  • 24 GB VRAM: no quant fits at an 8K context.
  • 32 GB VRAM: no quant fits at an 8K context.
  • 48 GB VRAM: no quant fits at an 8K context.

Which Kimi-K2.5 quant should you choose?

Q4_K_M is the usual starting point for Kimi-K2.5: it balances file size and quality well. Here it is a 621.2 GB file.

The Q8_0 file is 1091.1 GB, about 1.8 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

If memory is tight, the smallest file is TQ1_0 at 239.5 GB, with the quality trade-offs that come with fewer bits. Lower bit counts mean smaller files and more quality loss.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedeepseek2
Parameters1030B (approximate, mixture of experts)
Layers61
Trained context262,144 tokens
KV cache (fp16)130 KB per token

Source

Quantized byunsloth
FamilyKimi
Licenseother
Downloads (HF)13,177
Last updated2026-01-28

Frequently asked questions about Kimi-K2.5

How much VRAM does Kimi-K2.5 need?

About 622.8 GB at Q4_K_M with an 8K context window. That is the 621.2 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run Kimi-K2.5 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 241.1 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Kimi-K2.5 GGUF?

It depends on the quant: TQ1_0 is 239.5 GB, IQ1_S is 275.9 GB, IQ1_M is 300.6 GB, IQ2_XXS is 326.7 GB, IQ2_M is 345.2 GB, Q2_K is 373.8 GB and 21 more.

What context length does Kimi-K2.5 support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 1.1 GB at 8K and 4.3 GB at 32K.

Where can I download Kimi-K2.5 GGUF?

The files are on Hugging Face in the unsloth/Kimi-K2.5-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Kimi-K2.5 use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run Kimi-K2.5 on 8 GB VRAM?

No. The smallest quant requires at least 240 GB unified memory, far beyond 8 GB VRAM.

Does the model support vision tasks?

The current GGUF builds do not include vision support.

What is the recommended quant for 256 GB RAM and 16 GB VRAM?

Use Q4_K_S or Q4_K_M, which need roughly 585 GB unified memory.

Other GGUFs of Kimi-K2.5

Similar models

Browse all →

More from unsloth

All by unsloth →

More Kimi models

All Kimi →

Other large models

All large models →