What are the hardware requirements for FusionNet 34Bx2 MoE?

FusionNet 34Bx2 MoE requires approximately 121.2 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 and how large is it?

FusionNet 34Bx2 MoE is developed by TomGrc, 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.

How does FusionNet 34Bx2 MoE perform on standard benchmarks?

FusionNet 34Bx2 MoE has the following published scores: MMLU 77.05. Compare against reference models on this page or on the LLM Explorer leaderboards.

FusionNet 34Bx2 MoE by TomGrc — VRAM 121.2GB, 32K context

Name: FusionNet 34Bx2 MoE
Author: TomGrc

FusionNet 34Bx2 MoE is an open-source language model by TomGrc. Features: 60.8b LLM, VRAM: 121.2GB, Context: 32K, License: mit, MoE, LLM Explorer Score: 0.11, Arc: 73, HellaSwag: 86.2, MMLU: 77.1, GSM8K: 70.9.

Conversational En Endpoints compatible Mixtral Model-index Moe Region:us Safetensors Sharded Tensorflow

Model Card on HF 🤗: https://huggingface.co/TomGrc/FusionNet_34Bx2_MoE

FusionNet 34Bx2 MoE Parameters and Internals

Model Type

text generation

Additional Notes

Fine-tuned model on English language using the MoE method, enhancing its performance.

Training Details

Methodology:

MoE (Mixture of Experts)

Model Architecture:

MoE method to improve original model performance

LLM Name	FusionNet 34Bx2 MoE
Repository 🤗	https://huggingface.co/TomGrc/FusionNet_34Bx2_MoE
Model Size	60.8b
Required VRAM	121.2 GB
Updated	2026-07-23
Maintainer	TomGrc
Model Type	mixtral
Model Files	3.9 GB: 1-of-32 3.8 GB: 2-of-32 3.8 GB: 3-of-32 3.8 GB: 4-of-32 3.8 GB: 5-of-32 3.8 GB: 6-of-32 3.8 GB: 7-of-32 3.8 GB: 8-of-32 3.8 GB: 9-of-32 3.8 GB: 10-of-32 3.8 GB: 11-of-32 3.8 GB: 12-of-32 3.8 GB: 13-of-32 3.8 GB: 14-of-32 3.8 GB: 15-of-32 3.8 GB: 16-of-32 3.8 GB: 17-of-32 3.8 GB: 18-of-32 3.8 GB: 19-of-32 3.8 GB: 20-of-32 3.8 GB: 21-of-32 3.8 GB: 22-of-32 3.8 GB: 23-of-32 3.8 GB: 24-of-32 3.8 GB: 25-of-32 3.8 GB: 26-of-32 3.8 GB: 27-of-32 3.8 GB: 28-of-32 4.0 GB: 29-of-32 4.0 GB: 30-of-32 3.9 GB: 31-of-32 2.8 GB: 32-of-32
Supported Languages	en
Model Architecture	MixtralForCausalLM
License	mit
Context Length	32768
Model Max Length	32768
Transformers Version	4.36.2
Tokenizer Class	LlamaTokenizer
Padding Token	<s>
Vocabulary Size	64000
Torch Data Type	bfloat16

Rank the FusionNet 34Bx2 MoE 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

Quantized Models of the FusionNet 34Bx2 MoE

Model	Likes	Downloads	VRAM
FusionNet 34Bx2 MoE GGUF	5	355	22 GB
FusionNet 34Bx2 MoE AWQ	5	8	32 GB
FusionNet 34Bx2 MoE GPTQ	2	5	31 GB

Best Alternatives to FusionNet 34Bx2 MoE

Best Alternatives	Context / RAM	Downloads	Likes
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
... Cloudyu Mixtral 34Bx2 MoE 60B	195K / 121.8 GB	57	0
...DPO TomGrc FusionNet 34Bx2 MoE	32K / 121.8 GB	81	4
Nous Hermes 2 MoE 2x34B	4K / 121.9 GB	96	0

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