LG Releases K-EXAONE 2.0: A 750B Parameter Open MoE Model
Secure_Smoke_4280 · reddit · 2026-07-31
LG AI Research has officially released the K-EXAONE 2.0 model. The model features a massive 750B total parameters with 37B active parameters, utilizing a sparse Mixture-of-Experts (MoE) architecture.
The company has open-sourced multiple weight versions on Hugging Face, including the base model, FP8, NVFP4, and DSpark variants, catering to various compute and deployment requirements.
Related event: LG Releases K-EXAONE 2.0: A 750B Parameter Open-Source Model(4 posts)→
More from Models
- Reflection AI open-sources Beam under Apache 2.0 with 40+ ecosystem partners — zhyncs42 · 2026-10-16
- Why are sub-15B open models so rare now? Labs seem to have settled on 24-27B minimum — Hot_Example_4456 · 2026-10-16
- User spots model openly generating "ideal responses" to train other models — liminal_bardo · 2026-10-16
- Reflection releases open MoE Beam-1.0: 501B total/23B active, 1M context, 3x fewer reasoning tokens — jacek2023 · 2026-10-16
- Reflection's 501B MoE Beam hits 463 tok/s on 8x GB200 with NVFP4 at day 0 — zhyncs42 · 2026-10-16
- Gemma 4 31B QAT on one 24GB card: 46 tok/s plain, ~80 with draft model on RTX 4090 — Subject_Poet_3861 · 2026-10-16