LLM EXPLORER 56,604 MODELS INDEXED

Hermes 4 14B FP8 by NousResearch

By NousResearch · 15435 downloads
HF Repo ↗ Copy URL

Hermes 4 14B FP8 is an open-source language model by NousResearch. Features: 14b LLM, VRAM: 16.4GB, Context: 40K, License: apache-2.0, LLM Explorer Score: 0.25.

  Arxiv:2508.18255   Atropos Base model:quantized:qwen/qwen...   Base model:qwen/qwen3-14b   Chat   Chatml   Compressed-tensors   Conversational   Dataforge   En   Endpoints compatible   Finetuned   Function calling   Hybrid-mode   Instruct   Json mode   Long context   Qwen-3-14b   Qwen3   Reasoning   Region:us   Roleplaying   Safetensors   Sharded   Structured outputs   Tensorflow   Tool use
Model Card on HF 🤗: https://huggingface.co/NousResearch/Hermes-4-14B-FP8 

Hermes 4 14B FP8 Parameters and Internals

LLM NameHermes 4 14B FP8
Repository 🤗https://huggingface.co/NousResearch/Hermes-4-14B-FP8 
Base Model(s)  Qwen3 14B   Qwen/Qwen3-14B
Model Size14b
Required VRAM16.4 GB
Updated2026-07-19
MaintainerNousResearch
Model Typeqwen3
Model Files  4.9 GB: 1-of-4   5.0 GB: 2-of-4   4.9 GB: 3-of-4   1.6 GB: 4-of-4
Supported Languagesen
Model ArchitectureQwen3ForCausalLM
Licenseapache-2.0
Context Length40960
Model Max Length40960
Transformers Version4.52.4
Tokenizer ClassQwen2Tokenizer
Padding Token<|endoftext|>
Vocabulary Size151936
Torch Data Typebfloat16
Errorsreplace

Best Alternatives to Hermes 4 14B FP8

Best Alternatives
Context / RAM
Downloads
Likes
...JA Qwen3 14B Agentic 256K V0.1256K / 29.5 GB288
SimpleChat 14B V1195K / 29.5 GB102
...0528DistillQwen 14B V27.3 200K195K / 29.5 GB95
...uct 21B Brainstorm20x 128K Ctx128K / 84.1 GB80
NousCoder 14B80K / 29.5 GB379221
MiroThinker 14B DPO V0.264K / 29.7 GB286
UIGEN T3 14B Preview40K / 29.5 GB2422
Qwen3 14B40K / 29.7 GB2722973437
Hermes 4 14B40K / 29.5 GB194693169
Qwen3 14B NVFP440K / 10.6 GB12521413
Note: green Score (e.g. "73.2") means that the model is better than NousResearch/Hermes-4-14B-FP8.