Vision
Google says a lightweight embedding module replaces a conventional vision encoder, letting the language-model backbone perform visual processing.
Quick answer
Google released an encoder-free multimodal 12B model with a stated 16GB VRAM or unified-memory target, Apache 2.0 licensing, and support across local inference ecosystems.
Unified architecture
Google says a lightweight embedding module replaces a conventional vision encoder, letting the language-model backbone perform visual processing.
Google says raw audio is projected into the same dimensional space as text tokens without a separate audio encoder.
Text, vision, and audio meet inside a unified backbone, which is the defining encoder-free architecture claim for this release.
Start from Google's 16GB VRAM or unified-memory target, then budget for the runtime, quantization, KV cache, inputs, tools, and operating system.
Choose the exact pre-trained or instruction-tuned artifact and review its files, notices, and model card.
Google lists LM Studio, Ollama, AI Edge tools, Transformers, llama.cpp, MLX, SGLang, and vLLM as ecosystem paths; support and performance differ.
The launch says Apache 2.0. Review the exact distribution, dependencies, acceptable-use requirements, and deployment obligations.
Text, image and video models side by side. Compare them on the same prompt, pick one per shot, or let the agent choose.
Write it, picture it, film it. One sentence is enough to begin.