trillionlabs / Gravity-16B-A3B-Base

Text Generation16B2458Updated Sep 1, 2026

Downloads

245

Likes

8

Last updated

Sep 1, 2026

License

apache-2.0

Gravity-16B-A3B-Base is a pretrained language model trained from scratch (from random weights) by Trillion Labs, Lunit Consortium. Built on a sparse Mixture-of-Experts (MoE) architecture, it features 16.24B total parameters with 3.16B active parameters per token. The model was…

Tasks & library

  • text generation
  • transformers

Tags

  • transformers
  • gravity_moe
  • text-generation
  • moe
  • mixture-of-experts
  • gravity

Technical specifications

Parameters16B
Librarytransformers
Languagesen, ko
Licenseapache-2.0