What are the hardware requirements for FusionNet 34Bx2 MoE AWQ?

FusionNet 34Bx2 MoE AWQ requires approximately 32.8 GB of VRAM and supports a context window of 32K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed FusionNet 34Bx2 MoE AWQ and how large is it?

FusionNet 34Bx2 MoE AWQ is developed by TheBloke, a model with 60.8b 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 FusionNet 34Bx2 MoE AWQ?

FusionNet 34Bx2 MoE AWQ is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

FusionNet 34Bx2 MoE AWQ by TheBloke — VRAM 32.8GB, 32K context

Name: FusionNet 34Bx2 MoE AWQ
Author: TheBloke

FusionNet 34Bx2 MoE AWQ is an open-source language model by TheBloke. Features: 60.8b LLM, VRAM: 32.8GB, Context: 32K, License: mit, MoE, Quantized, LLM Explorer Score: 0.11.

4-bit Awq Base model:quantized:tomgrc/fu... Base model:tomgrc/fusionnet 34... Conversational En Mixtral Moe Quantized Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/TheBloke/FusionNet_34Bx2_MoE-AWQ

FusionNet 34Bx2 MoE AWQ Parameters and Internals

Model Type

mixtral

Use Cases

Areas:

text generation

Additional Notes

This model is tuned for MoE method which boosts performance significantly. For AutoAWQ inference, AutoAWQ 0.1.8 or later versions should be installed for compatibility.

Supported Languages

en (fine-tuned)

Training Details

Data Sources:

VMware Open Instruct

Methodology:

fine-tuned using MoE method

Context Length:

8192

Model Architecture:

FusionNet with 60.8B parameters, utilizing MoE (Mixture of Experts) method

Input Output

Input Format:

[INST] <> {system_message} <> {prompt} [/INST]

Accepted Modalities:

text

Output Format:

text generation

LLM Name	FusionNet 34Bx2 MoE AWQ
Repository 🤗	https://huggingface.co/TheBloke/FusionNet_34Bx2_MoE-AWQ
Model Name	FusionNet 34Bx2 MoE
Model Creator	Suqin Zhang
Base Model(s)	FusionNet 34Bx2 MoE TomGrc/FusionNet_34Bx2_MoE
Model Size	60.8b
Required VRAM	32.8 GB
Updated	2026-06-09
Maintainer	TheBloke
Model Type	mixtral
Model Files	9.9 GB: 1-of-4 9.9 GB: 2-of-4 9.9 GB: 3-of-4 3.1 GB: 4-of-4
Supported Languages	en
AWQ Quantization	Yes
Quantization Type	awq
Model Architecture	MixtralForCausalLM
License	mit
Context Length	32768
Model Max Length	32768
Transformers Version	4.37.0.dev0
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	64000
Torch Data Type	float16

Rank the FusionNet 34Bx2 MoE AWQ 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 FusionNet 34Bx2 MoE AWQ

Best Alternatives	Context / RAM	Downloads	Likes
Mixtral 34Bx2 MoE 60B AWQ	195K / 32.8 GB	7	4
...sionNet 34Bx2 MoE V0.1 DPO F16	195K / 121.8 GB	8097	14
60B MoE Coder V3	195K / 121.8 GB	70	3
Mixtral 34Bx2 MoE 60B	195K / 121.9 GB	7876	114
Yi 34Bx2 MoE 60B DPO	195K / 121.8 GB	8035	3
Bagel Hermes 2x34B	195K / 121.9 GB	70	16
Yi 34Bx2 MoE 200K	195K / 121.9 GB	7897	2
Yi 34Bx2 MoE 60B	195K / 121.9 GB	7954	64
...34Bx2 MoE V0.1 Full Linear DPO	195K / 121.8 GB	68	2
FusionNet 34Bx2 MoE V0.1	195K / 121.2 GB	52	8

Note: green Score (e.g. "73.2") means that the model is better than TheBloke/FusionNet_34Bx2_MoE-AWQ.