🚀 Frontier MoE — 128B–1T
Collection
Largest open-weight LLMs, BatiAI-quantized. Mac-runnable from M4 Max 128GB to Mac Studio M3 Ultra 512GB. • 7 items • Updated
DeepSeek-V4-Flash GGUF — superseded
This model has moved to batiai/DeepSeek-V4-Flash-0731-GGUF.
DeepSeek released a refreshed V4-Flash checkpoint on 2026-07-31. We quantized that revision instead — same architecture, newer weights, verified in Korean. The files that used to live here were removed to make room for it.
Use instead: https://huggingface.co/batiai/DeepSeek-V4-Flash-0731-GGUF
The original (2026-04) quants can be regenerated on request — open a discussion.
— BatiAI · on-device Korean AI · https://flow.bati.ai
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support