What are the hardware requirements for Kimi Dev 72B?

Kimi Dev 72B requires approximately 83.4 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 Dev 72B and how large is it?

Kimi Dev 72B is developed by moonshotai, a model with 72b 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 Dev 72B?

Kimi Dev 72B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

Kimi Dev 72B by moonshotai — VRAM 83.4GB, 128K context

Name: Kimi Dev 72B
Author: moonshotai

Kimi Dev 72B is an open-source language model by moonshotai. Features: 72b LLM, VRAM: 83.4GB, Context: 128K, License: mit, LLM Explorer Score: 0.23.

Base model:finetune:qwen/qwen2... Base model:qwen/qwen2.5-72b Code Conversational Endpoints compatible Issue-resolving Qwen2 Region:us Safetensors Sharded Software Swebench Tensorflow

Model Card on HF 🤗: https://huggingface.co/moonshotai/Kimi-Dev-72B

Kimi Dev 72B Parameters and Internals

LLM Name	Kimi Dev 72B
Repository 🤗	https://huggingface.co/moonshotai/Kimi-Dev-72B
Base Model(s)	Qwen/Qwen2.5-72B Qwen/Qwen2.5-72B
Model Size	72b
Required VRAM	83.4 GB
Updated	2026-07-22
Maintainer	moonshotai
Model Type	qwen2
Model Files	4.2 GB: 1-of-80 1.8 GB: 10-of-80 1.8 GB: 11-of-80 1.8 GB: 12-of-80 1.8 GB: 13-of-80 1.8 GB: 14-of-80 1.8 GB: 15-of-80 1.8 GB: 16-of-80 1.8 GB: 17-of-80 1.8 GB: 18-of-80 1.8 GB: 19-of-80 1.8 GB: 2-of-80 1.8 GB: 20-of-80 1.8 GB: 21-of-80 1.8 GB: 22-of-80 1.8 GB: 23-of-80 1.8 GB: 24-of-80 1.8 GB: 25-of-80 1.8 GB: 26-of-80 1.8 GB: 27-of-80 1.8 GB: 28-of-80 1.8 GB: 29-of-80 1.8 GB: 3-of-80 1.8 GB: 30-of-80 1.8 GB: 31-of-80 1.8 GB: 32-of-80 1.8 GB: 33-of-80 1.8 GB: 34-of-80 1.8 GB: 35-of-80 1.8 GB: 36-of-80 1.8 GB: 37-of-80 1.8 GB: 38-of-80 1.8 GB: 39-of-80 1.8 GB: 4-of-80 1.8 GB: 40-of-80 1.8 GB: 41-of-80 1.8 GB: 42-of-80 1.8 GB: 43-of-80 1.8 GB: 44-of-80 1.8 GB: 45-of-80 1.8 GB: 46-of-80 1.8 GB: 47-of-80 1.8 GB: 48-of-80 1.8 GB: 49-of-80 1.8 GB: 5-of-80
Model Architecture	Qwen2ForCausalLM
License	mit
Context Length	131072
Model Max Length	131072
Transformers Version	4.43.1
Vocabulary Size	152064
Torch Data Type	bfloat16

Rank the Kimi Dev 72B 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

Quantized Models of the Kimi Dev 72B

Model	Likes	Downloads	VRAM
Kimi Dev 72B 4bit DWQ	21	4465	40 GB
Kimi Dev 72B 8bit	2	361	77 GB
Kimi Dev 72B 5bit	2	368	50 GB
Kimi Dev 72B GPTQ 4bit	2	12	45 GB

Best Alternatives to Kimi Dev 72B

Best Alternatives	Context / RAM	Downloads	Likes
...R10528DistillQwen 72B Preview3	195K / 146 GB	8	7
KAT Dev 72B Exp FP8	156K / 75.5 GB	13	6
Qwen2.5 72B	128K / 145.5 GB	13708	101
Ultiima 72B	128K / 146.1 GB	51	3
Homer V1.0 Qwen2.5 72B	128K / 146.1 GB	14	6
Ultiima 72B V1.5	128K / 146.1 GB	25	0
EVA Qwen2.5 72B V0.2	128K / 146 GB	159	22
Qwen2 72B	128K / 145.5 GB	18480	199
AceInstruct 72B	128K / 146 GB	88	18
...n2.5 72B 2x Instruct TIES V1.0	128K / 146.1 GB	76	1

Note: green Score (e.g. "73.2") means that the model is better than moonshotai/Kimi-Dev-72B.