hugpadhugpad

Tech2wild / Nex-N2.5-mini-Int4-Int8Mix

Agents36B802Updated Sep 8, 2026
Hugging Face

Downloads

80

Likes

2

Last updated

Sep 8, 2026

License

apache-2.0

Data-free weight-only quantization of nex-agi/Nex-N2.5-mini (a Qwen3.5-35B-A3B-family agentic model with a vision tower, 262K context) in the compressed-tensors pack-quantized format that stock vLLM loads directly. 66 GB of BF16 becomes 20 GB, which fits a pair of RTX 3090s at…

Tasks & library

  • image text to text

Tags

  • qwen3_5_moe
  • compressed-tensors
  • int4
  • int8
  • agentic
  • vllm

Technical specifications

Parameters36B
LibraryNot available
LanguagesNot available
Base modelnex-agi/Nex-N2.5-mini
Licenseapache-2.0