header-langage
简体中文
繁體中文
English
Tiếng Việt
한국어
日本語
ภาษาไทย
Türkçe
Scan to Download the APP
Partnered With:

Google Open Sources Gemma 4 Model, Introduces 12B Version Targeting Laptops with 16GB RAM

According to Dongcha Beating monitoring, Google has released Gemma 4 12B, the first new model specification of the Gemma 4 family since its launch in April this year. Unlike the previous models covering ultra-lightweight edge side E2B, E4B, as well as the 31B, 26B-A4B aimed at high-performance inference, the new version is positioned for consumer-grade PC local running scenarios, focusing on deploying a multimodal agent on laptops with 16GB of memory.

Gemma 4 12B adopts a unified encoder-free multimodal architecture, supporting both text and image inputs. Google has simultaneously upgraded the LiteRT-LM local inference tool, adding OpenAI API compatibility service mode, allowing developers to directly connect tools such as Continue, Aider, and Open WebUI to the locally running Gemma 4 12B without relying on cloud-based models.

The entire Gemma 4 model family was officially released in April, so this is not a new generation model update but a product line extension. Previously, Gemma 4 had covered a parameter range of about 8B to 31B, with the 12B version occupying a gap position between performance and hardware thresholds. Instead of continuing to pursue larger parameter scales, Google is now more focused on enabling multimodal models with agent capabilities to run stably on regular laptops.

举报 Correction/Report
Correction/Report
Submit
Add Library
Visible to myself only
Public
Save
Choose Library
Add Library
Cancel
Finish