zenlm

Zen — open model family and training stack.

Zen

An open model family and the training stack behind it.

Models Open weights across sizes and modalities
Training Pretraining, fine-tuning, and evaluation pipelines
Serving Quantization and inference for local and hosted use

Start here

Claude Code conversation data collector for training
Updated 2026-07-26 19:58:37 +00:00
Enso Diffusion: Multimodal Mixture of Unbound Experts (MUEN) — unifying text, vision, and audio through diffusion-based intelligence.
Updated 2026-07-26 19:58:34 +00:00
Zen Musician 7B - Music generation from lyrics
Updated 2026-07-16 17:53:29 +00:00
Zen Guard - Content safety detection and moderation
Updated 2026-07-16 17:48:33 +00:00
🎹 Mu: Blazingly Fast and Embarrassingly Simple End-to-End Full-Length Song Generation with Latent Diffusion.
Updated 2026-07-11 07:11:03 +00:00
Zen Dub - Real-time high-quality lip-sync model
Updated 2026-07-11 07:10:04 +00:00
Updated 2026-07-08 01:31:10 +00:00
Zen Coder 4B - Advanced coding assistant
Updated 2026-06-29 04:02:59 +00:00
Real-time multimodal translation with voice cloning and lip synchronization
Updated 2026-06-29 04:00:21 +00:00
QLoRA training for Zen4 Ultra (1.04T MoE) — gate/router unfreezing approach for MoE uncensoring
Updated 2026-06-27 01:45:32 +00:00
QLoRA training space for Zen Max identity fine-tuning
Updated 2026-06-27 01:44:08 +00:00
Flagship Zen Coder Flash — 31B MoE model optimized for code generation
Updated 2026-06-27 01:43:47 +00:00
Unified fine-tuning for 100+ LLMs and VLMs
Updated 2026-06-26 08:54:54 +00:00
Zen Nano v1.0 - Ultra-lightweight edge AI model
Updated 2026-06-21 14:47:20 +00:00
Real-time speech translation for broadcast news monitoring. Powered by Hanzo AI.
Updated 2026-06-21 14:42:20 +00:00