Day Zero Gemma 4 Support on Overshoot

Younes El Hjouji

April 2, 2026

All Gemma 4 models available on Overshoot Playground, with 2 hours of free inference on sign up.

Day zero Gemma 4 launch. Google has rejoined the open model map with a forceful entry in the Gemma 4 family, dropped today on April 2, 2026.

All Gemma 4 models are already live on the Overshoot Playground, so if you want to skip the launch posts and just try them, start there.

The short version, focused on the part we care about: Gemma 4 is a new open model family from Google released under Apache 2.0, and all four models natively process images and video. Google also says the E2B and E4B variants add native audio input, the edge models carry a 128K context window, and the larger models go up to 256K. In their launch writeup they specifically call out visual tasks like OCR and chart understanding, which is the right place for an open multimodal family to start. Read the full launch post here.

The Models

Gemma 4 E2B
Gemma 4 E4B
Gemma 4 26B-A4B
Gemma 4 31B

The two small models are the obvious edge story. The 26B-A4B is the MoE latency play. The 31B is the dense quality play.

Gemma 4 benchmark chart from Google's launch post — Google's launch graphic showing Gemma 4's open-model chat arena positioning.

The Gemma 4 models are available on Overshoot for real-time inference on the day of their release. Try them out on our playground.

More to come: we'll put these through the same lens we use for the rest of the open model map: deployment behavior, latency, and where the vision capabilities actually hold up in practice.