What are the hardware requirements for Kimi K2 Thinking NVFP4?

Kimi K2 Thinking NVFP4 requires approximately 210 GB of VRAM and supports a context window of 256K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Kimi K2 Thinking NVFP4 and how large is it?

Kimi K2 Thinking NVFP4 is developed by nvidia. 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 Thinking NVFP4?

Kimi K2 Thinking NVFP4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Kimi K2 Thinking NVFP4 by nvidia — VRAM 210GB, 256K context

Name: Kimi K2 Thinking NVFP4
Author: nvidia

Kimi K2 Thinking NVFP4 is an open-source language model by nvidia. Features: LLM, VRAM: 210GB, Context: 256K, License: other, LLM Explorer Score: 0.27.

Base model:moonshotai/kimi-k2-... Base model:quantized:moonshota... Conversational Custom code Deepseek v3 Fp4 Kimi-k2 Model optimizer Modelopt Nvidia Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/nvidia/Kimi-K2-Thinking-NVFP4

Kimi K2 Thinking NVFP4 Parameters and Internals

LLM Name	Kimi K2 Thinking NVFP4
Repository 🤗	https://huggingface.co/nvidia/Kimi-K2-Thinking-NVFP4
Base Model(s)	Kimi K2 Thinking moonshotai/Kimi-K2-Thinking
Required VRAM	210 GB
Updated	2026-07-19
Maintainer	nvidia
Model Type	deepseek_v3
Model Files	5.0 GB: 1-of-119 5.0 GB: 2-of-119 5.0 GB: 3-of-119 5.0 GB: 4-of-119 5.0 GB: 5-of-119 5.0 GB: 6-of-119 5.0 GB: 7-of-119 5.0 GB: 8-of-119 5.0 GB: 9-of-119 5.0 GB: 10-of-119 5.0 GB: 11-of-119 5.0 GB: 12-of-119 5.0 GB: 13-of-119 5.0 GB: 14-of-119 5.0 GB: 15-of-119 5.0 GB: 16-of-119 5.0 GB: 17-of-119 5.0 GB: 18-of-119 5.0 GB: 19-of-119 5.0 GB: 20-of-119 5.0 GB: 21-of-119 5.0 GB: 22-of-119 5.0 GB: 23-of-119 5.0 GB: 24-of-119 5.0 GB: 25-of-119 5.0 GB: 26-of-119 5.0 GB: 27-of-119 5.0 GB: 28-of-119 5.0 GB: 29-of-119 5.0 GB: 30-of-119 5.0 GB: 31-of-119 5.0 GB: 32-of-119 5.0 GB: 33-of-119 5.0 GB: 34-of-119 5.0 GB: 35-of-119 5.0 GB: 36-of-119 5.0 GB: 37-of-119 5.0 GB: 38-of-119 5.0 GB: 39-of-119 5.0 GB: 40-of-119 5.0 GB: 41-of-119 5.0 GB: 42-of-119
Model Architecture	DeepseekV3ForCausalLM
License	other
Context Length	262144
Model Max Length	262144
Transformers Version	4.57.1
Vocabulary Size	163840

Rank the Kimi K2 Thinking NVFP4 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 Thinking NVFP4

Best Alternatives	Context / RAM	Downloads	Likes
Kimi K2 Thinking BF16	256K / 220 GB	63	6
Kimi K2 Instruct 0905	256K / 667.4 GB	17	7
DeepSeek R1 0528	160K / 176.1 GB	48	15
DeepSeek R1 0528 FP4	160K / 107 GB	72991	36
DeepSeek V3.0324 FP4	160K / 242.8 GB	9984	12
DeepSeek R1 FP4	160K / 232 GB	6856	264
DeepSeek V3.1 MLX 5.5bit	160K / GB	400	5
...V3.1 Base Int4 Mixed AutoRound	160K / 220 GB	103	2
...Seek V3.1 Int4 Mixed AutoRound	160K / 225 GB	53	11
R1 1776	160K / 243.4 GB	22	9

Note: green Score (e.g. "73.2") means that the model is better than nvidia/Kimi-K2-Thinking-NVFP4.