What are the hardware requirements for DeepSeek V4 Flash 0731 REAM128 146B Exp?

DeepSeek V4 Flash 0731 REAM128 146B Exp requires approximately 83.8 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 REAM128 146B Exp and how large is it?

DeepSeek V4 Flash 0731 REAM128 146B Exp is developed by WaveCut, a model with 180b 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 REAM128 146B Exp?

DeepSeek V4 Flash 0731 REAM128 146B Exp 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 REAM128 146B Exp by WaveCut — VRAM 83.8GB, 1024K context

Name: DeepSeek V4 Flash 0731 REAM128 146B Exp
Author: WaveCut

DeepSeek V4 Flash 0731 REAM128 146B Exp is an open-source language model by WaveCut. Features: 180b LLM, VRAM: 83.8GB, Context: 1024K, License: mit, LLM Explorer Score: 0.29.

8-bit Base model:quantized:wavecut/d... Base model:wavecut/deepseek-v4... Deepseek-v4 Deepseek v4 Endpoints compatible Fp4 Fp8 Mixture-of-experts Ream Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/WaveCut/DeepSeek-V4-Flash-0731-REAM128-146B-exp

DeepSeek V4 Flash 0731 REAM128 146B Exp Parameters and Internals

LLM Name	DeepSeek V4 Flash 0731 REAM128 146B Exp
Repository 🤗	https://huggingface.co/WaveCut/DeepSeek-V4-Flash-0731-REAM128-146B-exp
Base Model(s)	...eek V4 Flash 0731 REAM160 180B WaveCut/DeepSeek-V4-Flash-0731-REAM160-180B
Model Size	180b
Required VRAM	83.8 GB
Updated	2026-08-14
Maintainer	WaveCut
Model Type	deepseek_v4
Model Files	2.1 GB: 1-of-44 1.9 GB: 2-of-44 1.9 GB: 3-of-44 1.9 GB: 4-of-44 1.9 GB: 5-of-44 1.9 GB: 6-of-44 1.9 GB: 7-of-44 1.9 GB: 8-of-44 1.9 GB: 9-of-44 1.9 GB: 10-of-44 1.9 GB: 11-of-44 1.9 GB: 12-of-44 1.9 GB: 13-of-44 1.9 GB: 14-of-44 1.9 GB: 15-of-44 1.9 GB: 16-of-44 1.9 GB: 17-of-44 1.9 GB: 18-of-44 1.9 GB: 19-of-44 1.9 GB: 20-of-44 1.9 GB: 21-of-44 1.9 GB: 22-of-44 1.9 GB: 23-of-44 1.9 GB: 24-of-44 1.9 GB: 25-of-44 1.9 GB: 26-of-44 1.9 GB: 27-of-44 1.9 GB: 28-of-44 1.9 GB: 29-of-44 1.9 GB: 30-of-44 1.9 GB: 31-of-44 1.9 GB: 32-of-44 1.9 GB: 33-of-44 1.9 GB: 34-of-44 1.9 GB: 35-of-44 1.9 GB: 36-of-44 1.9 GB: 37-of-44 1.9 GB: 38-of-44 1.9 GB: 39-of-44 1.9 GB: 40-of-44 1.9 GB: 41-of-44 1.9 GB: 42-of-44 1.9 GB: 43-of-44 1.9 GB: 44-of-44
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 REAM128 146B 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 DeepSeek V4 Flash 0731 REAM128 146B Exp

Best Alternatives	Context / RAM	Downloads	Likes
DeepSeek V4 Flash 180B	1024K / 93.1 GB	10492	31
...eek V4 Flash 0731 REAM160 180B	1024K / 101 GB	15	0

Note: green Score (e.g. "73.2") means that the model is better than WaveCut/DeepSeek-V4-Flash-0731-REAM128-146B-exp.