Gemma 4 E2B on iPhone

What is Gemma 4 E2B?

Gemma 4 is Google’s latest generation of open models, built from the research behind Gemini. The E2B variant is sized for devices, with clearly stronger reasoning than Gemma 3 1B. In Priobs it runs as a 4-bit quantized MLX build, and at 3.6 GB it is the largest download in the library.

Who made it

Google releases Gemma models under the Gemma licence, which lets apps like Priobs run them locally. Gemma 4 is the newest generation, and it runs without ever touching Google’s servers.

What it’s good at on iPhone

Gemma 4 E2B is the pick for Google’s latest model, run locally:

• Multi-step reasoning and nuanced questions
• Longer writing that needs good context tracking
• Anyone who wants Google’s newest open model without the cloud

Tool support for Gemma 4 E2B is in beta: it can use Priobs tools like calendar and reminders, but expect the odd miss.

It asks more of your iPhone in speed, battery and memory. If you want the library’s most capable all-rounder with solid tool support, Qwen 3 4B is the better pick and a smaller download.

Which iPhones can run it

Gemma 4 E2B needs at least 4.5 GB of memory. Priobs checks your iPhone, hides models it does not have the memory for, and warns you if a model is close to the limit and may be unstable. That makes it the most demanding model in the library. As a rough guide, an iPhone 15 Pro or newer (8 GB of RAM) is the safe bet. On 6 GB iPhones it may not appear, or Priobs may mark it as possibly unstable.