LLM EXPLORER 65,548 MODELS INDEXED

CognitiveFusion2 4x7B BF16 by Kquant03

By Kquant03 · 8354 downloads
HF Repo ↗ Copy URL

CognitiveFusion2 4x7B BF16 is an open-source language model by Kquant03. Features: 24.2b LLM, VRAM: 48.3GB, Context: 32K, License: apache-2.0, MoE, LLM Explorer Score: 0.19, Arc: 73.4, HellaSwag: 89.2, MMLU: 64.3, GSM8K: 71.3.

  Arxiv:2101.03961   En   Endpoints compatible   Merge   Mixtral   Moe   Region:us   Safetensors   Sharded   Tensorflow
Model Card on HF 🤗: https://huggingface.co/Kquant03/CognitiveFusion2-4x7B-BF16 

CognitiveFusion2 4x7B BF16 Parameters and Internals

Model Type 
MoE, Transformer
Additional Notes 
This model employs the Mixture of Experts (MoE) framework, which enables efficient pretraining and faster inference by utilizing specific experts in its network.
Supported Languages 
en (Proficient)
LLM NameCognitiveFusion2 4x7B BF16
Repository 🤗https://huggingface.co/Kquant03/CognitiveFusion2-4x7B-BF16 
Model Size24.2b
Required VRAM48.3 GB
Updated2026-06-06
MaintainerKquant03
Model Typemixtral
Model Files  9.9 GB: 1-of-5   10.0 GB: 2-of-5   10.0 GB: 3-of-5   10.0 GB: 4-of-5   8.4 GB: 5-of-5
Supported Languagesen
Model ArchitectureMixtralForCausalLM
Licenseapache-2.0
Context Length32768
Model Max Length32768
Transformers Version4.38.2
Tokenizer ClassLlamaTokenizer
Padding Token<s>
Vocabulary Size32000
Torch Data Typebfloat16

Best Alternatives to CognitiveFusion2 4x7B BF16

Best Alternatives
Context / RAM
Downloads
Likes
Dzakwan MoE 4x7b Beta32K / 48.4 GB90720
Beyonder 4x7B V332K / 48.3 GB815660
Calme 4x7B MoE V0.232K / 48.3 GB82582
Calme 4x7B MoE V0.132K / 48.3 GB80372
Mera Mix 4x7B32K / 48.3 GB785519
MixtureofMerges MoE 4x7b V532K / 48.3 GB77151
MixtureofMerges MoE 4x7b V432K / 48.3 GB78384
Proto Athena 4x7B32K / 48.4 GB50
Proto Athena V0.2 4x7B32K / 48.4 GB80
NeuralStar FusionWriter 4x7b32K / 48.3 GB499
Note: green Score (e.g. "73.2") means that the model is better than Kquant03/CognitiveFusion2-4x7B-BF16.