What are the hardware requirements for DeepSeek V4 Flash 0731 JANG?

DeepSeek V4 Flash 0731 JANG requires approximately 37 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 JANG and how large is it?

DeepSeek V4 Flash 0731 JANG is developed by OsaurusAI, a model with 31.4b 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 DeepSeek V4 Flash 0731 JANG?

DeepSeek V4 Flash 0731 JANG 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 JANG by OsaurusAI — VRAM 37GB, 1024K context

Name: DeepSeek V4 Flash 0731 JANG
Author: OsaurusAI

DeepSeek V4 Flash 0731 JANG is an open-source language model by OsaurusAI. Features: 31.4b LLM, VRAM: 37GB, Context: 1024K, License: mit, Quantized, LLM Explorer Score: 0.35.

Affine Apple-silicon Awq Base model:deepseek-ai/deepsee... Base model:quantized:deepseek-... Deepseek Deepseek-v4 Deepseek-v4-flash-0731 Deepseek v4 Dsml Gptq Jang Mixture-of-experts Mla Mlx Osaurus Qat Quantized Reasoning Region:us Safetensors Sharded Tensorflow Tool-use

Model Card on HF 🤗: https://huggingface.co/OsaurusAI/DeepSeek-V4-Flash-0731-JANG

DeepSeek V4 Flash 0731 JANG Parameters and Internals

LLM Name	DeepSeek V4 Flash 0731 JANG
Repository 🤗	https://huggingface.co/OsaurusAI/DeepSeek-V4-Flash-0731-JANG
Base Model(s)	DeepSeek V4 Flash 0731 deepseek-ai/DeepSeek-V4-Flash-0731
Model Size	31.4b
Required VRAM	37 GB
Updated	2026-08-09
Maintainer	OsaurusAI
Model Type	deepseek_v4
Model Files	0.0 GB 1.0 GB: 1-of-102 1.0 GB: 2-of-102 1.0 GB: 3-of-102 1.0 GB: 4-of-102 1.0 GB: 5-of-102 1.0 GB: 6-of-102 1.0 GB: 7-of-102 1.0 GB: 8-of-102 1.0 GB: 9-of-102 1.0 GB: 10-of-102 1.0 GB: 11-of-102 1.0 GB: 12-of-102 1.0 GB: 13-of-102 1.0 GB: 14-of-102 1.0 GB: 15-of-102 1.0 GB: 16-of-102 1.0 GB: 17-of-102 1.0 GB: 18-of-102 1.0 GB: 19-of-102 1.0 GB: 20-of-102 1.0 GB: 21-of-102 1.0 GB: 22-of-102 1.0 GB: 23-of-102 1.0 GB: 24-of-102 1.0 GB: 25-of-102 1.0 GB: 26-of-102 1.0 GB: 27-of-102 1.0 GB: 28-of-102 1.0 GB: 29-of-102 1.0 GB: 30-of-102 1.0 GB: 31-of-102 1.0 GB: 32-of-102 1.0 GB: 33-of-102 1.0 GB: 34-of-102 1.0 GB: 35-of-102 1.0 GB: 36-of-102 1.0 GB: 37-of-102
GPTQ Quantization	Yes
AWQ Quantization	Yes
Quantization Type	gptq\|awq
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 JANG 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