What are the hardware requirements for DeepSeek V4 Flash 0731 Abliterated NVFP4?

DeepSeek V4 Flash 0731 Abliterated NVFP4 requires approximately 145.5 GB of VRAM and supports a context window of 1024K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed DeepSeek V4 Flash 0731 Abliterated NVFP4 and how large is it?

DeepSeek V4 Flash 0731 Abliterated NVFP4 is developed by sakamakismile, a model with 304.2b 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 DeepSeek V4 Flash 0731 Abliterated NVFP4?

DeepSeek V4 Flash 0731 Abliterated NVFP4 is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

DeepSeek V4 Flash 0731 Abliterated NVFP4 by sakamakismile — VRAM 145.5GB, 1024K context

Name: DeepSeek V4 Flash 0731 Abliterated NVFP4
Author: sakamakismile

DeepSeek V4 Flash 0731 Abliterated NVFP4 is an open-source language model by sakamakismile. Features: 304.2b LLM, VRAM: 145.5GB, Context: 1024K, License: mit, LLM Explorer Score: 0.34.

8-bit Abliterated Base model:apetersson/deepseek... Base model:quantized:apetersso... Deepseek-v4 Deepseek v4 Dspark Endpoints compatible Fp8 Mixture-of-experts Model-surgery Modelopt Nvfp4 Region:us Safetensors Sharded Tensorflow Vllm

Model Card on HF 🤗: https://huggingface.co/sakamakismile/DeepSeek-V4-Flash-0731-Abliterated-NVFP4

DeepSeek V4 Flash 0731 Abliterated NVFP4 Parameters and Internals

LLM Name	DeepSeek V4 Flash 0731 Abliterated NVFP4
Repository 🤗	https://huggingface.co/sakamakismile/DeepSeek-V4-Flash-0731-Abliterated-NVFP4
Base Model(s)	DeepSeek V4 Flash 0731 ... V4 Flash 0731 Abliterated FP8 deepseek-ai/DeepSeek-V4-Flash-0731 apetersson/DeepSeek-V4-Flash-0731-Abliterated-FP8
Model Size	304.2b
Required VRAM	145.5 GB
Updated	2026-08-11
Maintainer	sakamakismile
Model Type	deepseek_v4
Model Files	1.1 GB: 1-of-48 3.8 GB: 2-of-48 3.8 GB: 3-of-48 3.8 GB: 4-of-48 3.8 GB: 5-of-48 3.8 GB: 6-of-48 3.8 GB: 7-of-48 3.8 GB: 8-of-48 3.8 GB: 9-of-48 3.8 GB: 10-of-48 3.8 GB: 11-of-48 3.8 GB: 12-of-48 3.8 GB: 13-of-48 3.8 GB: 14-of-48 3.8 GB: 15-of-48 3.8 GB: 16-of-48 3.8 GB: 17-of-48 3.8 GB: 18-of-48 3.8 GB: 19-of-48 3.8 GB: 20-of-48 3.8 GB: 21-of-48 3.8 GB: 22-of-48 3.8 GB: 23-of-48 3.8 GB: 24-of-48 3.8 GB: 25-of-48 3.8 GB: 26-of-48 3.8 GB: 27-of-48 3.8 GB: 28-of-48 3.8 GB: 29-of-48 3.8 GB: 30-of-48 3.8 GB: 31-of-48 3.8 GB: 32-of-48 3.8 GB: 33-of-48 3.8 GB: 34-of-48 3.8 GB: 35-of-48 3.8 GB: 36-of-48 3.8 GB: 37-of-48 3.8 GB: 38-of-48 3.8 GB: 39-of-48
Model Architecture	DeepseekV4ForCausalLM
License	mit
Context Length	1048576
Model Max Length	1048576
Transformers Version	4.57.1
Vocabulary Size	129280
Torch Data Type	bfloat16

Rank the DeepSeek V4 Flash 0731 Abliterated 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 DeepSeek V4 Flash 0731 Abliterated NVFP4

Best Alternatives	Context / RAM	Downloads	Likes
DeepSeek V4 Flash 0731	1024K / 152.3 GB	785771	2803
... V4 Flash 0731 Abliterated FP8	1024K / 145.1 GB	5149	35
DeepSeek V4 Flash 0731 NVFP4	1024K / 149.3 GB	7421	6
DeepSeek V4 Flash 0731	1024K / 148.7 GB	89	25
DeepSeek V4 Flash 0731 NVFP4	1024K / 156.9 GB	75	2

Note: green Score (e.g. "73.2") means that the model is better than sakamakismile/DeepSeek-V4-Flash-0731-Abliterated-NVFP4.