KI

Kimi Linear REAP 35B A3B Instruct i1

KimiGGUF by mradermacherUpdated 2026-02-16
Download on Hugging Face
Parameters35B
Context1024K
Layers27
Architecturekimi-linear
Smallest quant0.1 GB
Quick answer

Kimi Linear REAP 35B A3B Instruct i1 needs about 30.9 GB of memory at Q4_K_M (a 21.3 GB file) with an 8K context window, so it fits on a 32 GB GPU. This repository has 24 quants, from 0.1 GB to 28.9 GB (Q6_K). Longer context needs more: at 32K tokens the KV cache alone adds about 36.7 GB.

Kimi Linear REAP 35B A3B Instruct i1 – a 35B sparse Mixture‑of‑Experts model compressed from 48B

Kimi Linear REAP 35B A3B Instruct i1 is a 35B parameter sparse Mixture‑of‑Experts causal language model. It retains near‑lossless performance compared to the full 48B model while being 30% lighter. The model supports a context length of 1,048,576 tokens and is released under the MIT license.

Good at

  • Code generation
  • Math & reasoning
  • Long‑context question answering

Limits

  • Large VRAM requirement for higher‑quality quants
  • No vision support
  • Requires GPU for efficient inference

Best for

  • Local deployment on resource‑constrained machines
  • Academic research with limited GPU budgets
  • Fine‑tuning on specialized datasets

Who it’s for. Developers and researchers who need a high‑performance language model that fits into tighter memory budgets. The model is suitable for local inference on GPUs with moderate VRAM and for use cases that demand long‑context understanding.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—0.1 GB 9.8 GB
Too big Get file →
IQ1_S7.3 GB 17.0 GB
Too big Get file →
IQ1_M8.1 GB 17.8 GB
Too big Get file →
IQ2_XXS9.4 GB 19.1 GB
Too big Get file →
IQ2_XS10.5 GB 20.1 GB
Too big Get file →
IQ2_S10.6 GB 20.2 GB
Too big Get file →
IQ2_M11.6 GB 21.3 GB
Too big Get file →
Q2_K_S12.0 GB 21.7 GB
Too big Get file →
Q2_K12.9 GB 22.6 GB
Too big Get file →
IQ3_XXS13.6 GB 23.3 GB
Too big Get file →
IQ3_XS14.5 GB 24.1 GB
Too big Get file →
IQ3_S15.3 GB 25.0 GB
Too big Get file →
Q3_K_S15.3 GB 25.0 GB
Too big Get file →
IQ3_M15.5 GB 25.1 GB
Too big Get file →
Q3_K_M16.8 GB 26.5 GB
Too big Get file →
Q3_K_L18.3 GB 28.0 GB
Too big Get file →
IQ4_XS18.8 GB 28.5 GB
Too big Get file →
Q4_020.0 GB 29.6 GB
Too big Get file →
Q4_K_S20.0 GB 29.7 GB
Too big Get file →
Q4_K_M21.3 GB 30.9 GB
Too big Get file →
Q4_122.1 GB 31.8 GB
Too big Get file →
Q5_K_S24.3 GB 33.9 GB
Too big Get file →
Q5_K_M24.9 GB 34.6 GB
Too big Get file →
Q6_K28.9 GB 38.5 GB
Too big Get file →

For 8 GB VRAM, use IQ1_S (7.31 GB) or IQ1_M (8.09 GB). If 12 GB is available, Q2_K (12.95 GB) or IQ2_XS (10.45 GB) provide better quality. For 16 GB or more, Q3_K_M (16.84 GB) or Q4_K_M (21.26 GB) offer higher fidelity. Memory estimates are approximate.

What GPU do you need for Kimi Linear REAP 35B A3B Instruct i1?

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: the best quant that fits is the main file (0.1 GB file, about 9.8 GB in use).
  • 16 GB VRAM: the best quant that fits is the main file (0.1 GB file, about 9.8 GB in use).
  • 24 GB VRAM: the best quant that fits is IQ3_XXS (13.6 GB file, about 23.3 GB in use).
  • 32 GB VRAM: the best quant that fits is Q4_1 (22.1 GB file, about 31.8 GB in use).
  • 48 GB VRAM: the best quant that fits is Q6_K (28.9 GB file, about 38.5 GB in use).

Which Kimi Linear REAP 35B A3B Instruct i1 quant should you choose?

Q4_K_M is the usual starting point for Kimi Linear REAP 35B A3B Instruct i1: it balances file size and quality well. Here it is a 21.3 GB file.

If memory is tight, the smallest file is the first one in the table at 0.1 GB, with the quality trade-offs that come with fewer bits.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturekimi-linear
Parameters35B (approximate, mixture of experts)
Active parameters3B
Layers27
Trained context1,048,576 tokens
KV cache (fp16)1094 KB per token

Source

Quantized bymradermacher
FamilyKimi
Licensemit
Downloads (HF)1,873
Last updated2026-02-16

Frequently asked questions about Kimi Linear REAP 35B A3B Instruct i1

How much VRAM does Kimi Linear REAP 35B A3B Instruct i1 need?

About 30.9 GB at Q4_K_M with an 8K context window. That is the 21.3 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 Linear REAP 35B A3B Instruct i1 on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 9.8 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Kimi Linear REAP 35B A3B Instruct i1 GGUF?

It depends on the quant: file is 0.1 GB, IQ1_S is 7.3 GB, IQ1_M is 8.1 GB, IQ2_XXS is 9.4 GB, IQ2_XS is 10.5 GB, IQ2_S is 10.6 GB and 18 more.

What context length does Kimi Linear REAP 35B A3B Instruct i1 support?

The GGUF header lists a trained context length of 1,048,576 tokens (1024K). The KV cache grows with the context you use: about 9.2 GB at 8K and 36.7 GB at 32K.

Where can I download Kimi Linear REAP 35B A3B Instruct i1 GGUF?

The files are on Hugging Face in the mradermacher/Kimi-Linear-REAP-35B-A3B-Instruct-i1-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 Linear REAP 35B A3B Instruct i1 use?

Hugging Face lists the license as mit. 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 this model on 8 GB VRAM?

Yes, using the IQ1_S or IQ1_M quant files, which are 7.31 GB and 8.09 GB respectively. The estimated VRAM usage at 8k context is 17 GB and 17.8 GB.

Does this model support vision tasks?

No, the model is a text‑only causal language model and does not include vision capabilities.

What license does it use?

The model is released under the MIT license, allowing free use and modification.

Other GGUFs of Kimi-Linear-REAP-35B-A3B-Instruct

Similar models

Browse all →

More from mradermacher

All by mradermacher →

More Kimi models

All Kimi →

Other large models

All large models →