Google's new open source Gemma 4 12B analyzes audio, video — and runs entirely locally on a typical 16GB enterprise laptop - VentureBeat
Quick Answer
Google has launched Gemma 4 12B, an open-source model with 11.95 billion parameters, optimized to run locally on a standard 16GB enterprise laptop.
Quick Take
Google has launched Gemma 4 12B, an open-source model with 11.95 billion parameters, optimized to run locally on a standard 16GB enterprise laptop. Its innovative 'Unified' architecture allows direct processing of audio and video without additional latency, making it a significant advancement in local AI capabilities.
Key Points
- Gemma 4 12B features 11.95 billion parameters and runs on 16GB VRAM.
- The model is released under a permissive Apache 2.0 license.
- It utilizes a unique encoder-free 'Unified' architecture for efficiency.
- Direct processing of audio and video reduces latency and memory usage.
- Available for immediate use, targeting local AI applications.
Article Excerpt
From source RSS / original summary# Google's new open source Gemma 4 12B analyzes audio, video — and runs entirely locally on a typical 16GB enterprise laptop. While many AI open source model providers are pursuing larger and more powerful models, Google is still giving attention to the smaller, more local side of the market. Today, the tech giant released Gemma 4 12B, an 11. 95-billion-parameter open-weights model with permissive Apache 2.
0 license optimized to execute locally on a standard enterprise laptop using just 16GB of VRAM or unified memory. Gemma 4 12B's most notable breakthrough is an encoder-free "Unified" architecture, which allows raw audio waveforms and visual patches to flow directly into the core LLM backbone without the latency or memory overhead of secondary processing modules. Available immediately for
Reader Mode unavailable (could not extract clean content).
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from WebSearch (Tavily)
See more →Microsoft debuts Scout agent, homegrown reasoning model - Axios
Microsoft introduced its first internally developed reasoning model, MAI-Thinking-1, with 35 billion parameters, and a personal AI agent named Scout at the Build conference. Scout will integrate with tools like Outlook and Teams to assist with tasks such as meeting preparation, while Microsoft also unveiled updates to its coding, voice, and image models.