Google has introduced Gemma 4, its latest family of open AI models designed for advanced reasoning and agent-based workflows.
Gemma 4 builds on the company’s earlier open models and is released under an Apache 2.0 license, allowing commercial use and modification.
According to Google, the models have already seen strong developer interest, with previous Gemma versions reaching over 400 million downloads.
The new lineup includes four model sizes:
The larger models are aimed at high-performance tasks, while the smaller versions are optimized for mobile and edge devices.
Google says the 31B model ranks among the top open models on the Arena AI leaderboard, with the 26B variant also placing in the top 10.
Both are designed to deliver strong performance while using less hardware compared to larger models.
Gemma 4 introduces improvements in multi-step reasoning, coding, and instruction-following. It also supports agentic workflows, including function calling and structured outputs, enabling developers to build tools that can interact with APIs and execute tasks.
All models support multimodal inputs such as images and video, while the smaller E2B and E4B versions also include audio input for speech-related tasks. The context length has been extended to 256K tokens for larger models and 128K for edge models.
The models are trained to support more than 140 languages and are designed to run across a wide range of hardware, from smartphones and laptops to GPUs and cloud infrastructure.
Gemma 4 is available through platforms such as Google AI Studio, Hugging Face, and other developer tools, with support for local and cloud-based deployment.
ManilaShaker is a tech media producing insightful and helpful content for our local and growing international audience. Our goal is to create a premier Philippine digital consumer electronics resource that provides the most objective reviews and comparisons globally.
