Skip to content
AI HOT (Curated Pool)

Google DeepMind Releases Gemma 4 12B, a Unified Encoder-Free Multimodal Model

Google DeepMind 发布 Gemma 4 12B:统一的无编码器多模态模型

Google DeepMind released Gemma 4 12B, a multimodal model with a unified encoder-free architecture, native audio input, Apache 2.0 licensing, and local laptop runtime with 16GB of VRAM or unified memory.

Why it matters: Google DeepMind open-sourced Gemma 4 12B, dropping separate vision and audio encoders so one Transformer handles images, text and speech, and it runs on 16GB of VRAM.

Read the original ↗Export Markdown