Google Releases Gemma 4 12B, a Free AI Model That Runs on Laptops
Gemma 4 12B brings a high-performance, multimodal AI model to consumer laptops without requiring cloud access, and its direct audio processing reduces latency and memory use.
Reporting from 1 sources: GIGAZINE.
Google released the Gemma 4 12B open AI model on June 3, 2026. The model runs on PCs with 16GB of VRAM or unified memory, supports image and audio input natively, and is free under the Apache License 2.0. It fills the gap between Google's edge-device models and its high-performance PC models.
Google released the Gemma 4 12B open model on June 3, filling the gap between the edge-device E4B and the high-performance 26B A4B models released in April. The model requires 16GB of VRAM or unified memory, putting it within reach of many laptops. It processes images and audio directly without an encoder, cutting latency and memory consumption. In benchmarks, Gemma 4 12B scored close to the larger 26B A4B and surpassed it in some tests. The model is free under the Apache License 2.0 and is already supported by LM Studio and ollama.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.