What are the hardware requirements for Kimi K2 Instruct 4bit?

Kimi K2 Instruct 4bit requires approximately 143.2 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Kimi K2 Instruct 4bit and how large is it?

Kimi K2 Instruct 4bit is developed by mlx-community, a model with 1026.4b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate Kimi K2 Instruct 4bit?

Kimi K2 Instruct 4bit is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Kimi K2 Instruct 4bit by mlx-community — VRAM 143.2GB, 128K context

Name: Kimi K2 Instruct 4bit
Author: mlx-community

Kimi K2 Instruct 4bit is an open-source language model by mlx-community. Features: 1026.4b LLM, VRAM: 143.2GB, Context: 128K, License: other, Quantized, Instruction-Based, LLM Explorer Score: 0.22.

4-bit 4bit Base model:moonshotai/kimi-k2-... Base model:quantized:moonshota... Conversational Custom code Instruct Kimi k2 Mlx Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/mlx-community/Kimi-K2-Instruct-4bit

Kimi K2 Instruct 4bit Parameters and Internals

LLM Name	Kimi K2 Instruct 4bit
Repository 🤗	https://huggingface.co/mlx-community/Kimi-K2-Instruct-4bit
Base Model(s)	Kimi K2 Instruct moonshotai/Kimi-K2-Instruct
Model Size	1026.4b
Required VRAM	143.2 GB
Updated	2026-07-17
Maintainer	mlx-community
Model Type	kimi_k2
Instruction-Based	Yes
Model Files	4.2 GB: 1-of-180 3.2 GB: 2-of-180 3.3 GB: 3-of-180 3.2 GB: 4-of-180 3.2 GB: 5-of-180 3.3 GB: 6-of-180 3.2 GB: 7-of-180 3.2 GB: 8-of-180 3.3 GB: 9-of-180 3.2 GB: 10-of-180 3.2 GB: 11-of-180 3.3 GB: 12-of-180 3.2 GB: 13-of-180 3.2 GB: 14-of-180 3.3 GB: 15-of-180 3.2 GB: 16-of-180 3.2 GB: 17-of-180 3.3 GB: 18-of-180 3.2 GB: 19-of-180 3.2 GB: 20-of-180 3.3 GB: 21-of-180 3.2 GB: 22-of-180 3.2 GB: 23-of-180 3.3 GB: 24-of-180 3.2 GB: 25-of-180 3.2 GB: 26-of-180 3.3 GB: 27-of-180 3.2 GB: 28-of-180 3.2 GB: 29-of-180 3.3 GB: 30-of-180 3.2 GB: 31-of-180 3.2 GB: 32-of-180 3.3 GB: 33-of-180 3.2 GB: 34-of-180 3.2 GB: 35-of-180 3.3 GB: 36-of-180 3.2 GB: 37-of-180 3.2 GB: 38-of-180 3.3 GB: 39-of-180 3.2 GB: 40-of-180 3.2 GB: 41-of-180 3.3 GB: 42-of-180 3.2 GB: 43-of-180 3.2 GB: 44-of-180
Quantization Type	4bit
Model Architecture	DeepseekV3ForCausalLM
License	other
Context Length	131072
Model Max Length	131072
Transformers Version	4.48.3
Vocabulary Size	163840
Torch Data Type	bfloat16

Rank the Kimi K2 Instruct 4bit capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Kimi K2 Instruct 4bit

Best Alternatives	Context / RAM	Downloads	Likes
...i K2 Instruct 0905 Mlx DQ3 K M	256K / 204.6 GB	2234	12
Kimi K2 Instruct 0905 BF16	256K / 1399.1 GB	31	4
Kimi K2 Instruct BF16	128K / 1367.3 GB	20	4

Note: green Score (e.g. "73.2") means that the model is better than mlx-community/Kimi-K2-Instruct-4bit.