What are the hardware requirements for KAT Dev 72B Exp?

KAT Dev 72B Exp requires approximately 146 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 KAT Dev 72B Exp and how large is it?

KAT Dev 72B Exp is developed by Kwaipilot, 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 KAT Dev 72B Exp?

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

KAT Dev 72B Exp by Kwaipilot — VRAM 146GB, 128K context

Name: KAT Dev 72B Exp
Author: Kwaipilot

KAT Dev 72B Exp is an open-source language model by Kwaipilot. Features: 72b LLM, VRAM: 146GB, Context: 128K, License: apache-2.0, LLM Explorer Score: 0.23.

Conversational Endpoints compatible Qwen2 Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/Kwaipilot/KAT-Dev-72B-Exp

KAT Dev 72B Exp Parameters and Internals

LLM Name	KAT Dev 72B Exp
Repository 🤗	https://huggingface.co/Kwaipilot/KAT-Dev-72B-Exp
Model Size	72b
Required VRAM	146 GB
Updated	2026-08-08
Maintainer	Kwaipilot
Model Type	qwen2
Model Files	4.5 GB: 1-of-31 5.0 GB: 2-of-31 4.8 GB: 3-of-31 4.8 GB: 4-of-31 4.8 GB: 5-of-31 5.0 GB: 6-of-31 4.8 GB: 7-of-31 4.8 GB: 8-of-31 4.8 GB: 9-of-31 5.0 GB: 10-of-31 4.8 GB: 11-of-31 4.8 GB: 12-of-31 4.8 GB: 13-of-31 5.0 GB: 14-of-31 4.8 GB: 15-of-31 4.8 GB: 16-of-31 4.8 GB: 17-of-31 5.0 GB: 18-of-31 4.8 GB: 19-of-31 4.8 GB: 20-of-31 4.8 GB: 21-of-31 5.0 GB: 22-of-31 4.8 GB: 23-of-31 4.8 GB: 24-of-31 4.8 GB: 25-of-31 5.0 GB: 26-of-31 4.8 GB: 27-of-31 4.8 GB: 28-of-31 4.8 GB: 29-of-31 3.2 GB: 30-of-31 2.5 GB: 31-of-31
Model Architecture	Qwen2ForCausalLM
License	apache-2.0
Context Length	131072
Model Max Length	131072
Transformers Version	4.51.1
Tokenizer Class	Qwen2Tokenizer
Padding Token	<\|endoftext\|>
Vocabulary Size	152064
Torch Data Type	bfloat16
Errors	replace

Rank the KAT Dev 72B Exp 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 KAT Dev 72B Exp

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	55	1

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