Gemma 4 Is Here: Run It on Featherless.

Google has released Gemma 4, a 31B model that significantly outperforms its predecessor, Gemma 3, and ranks fourth on the Arena AI leaderboard despite its relatively modest size compared to other large models. Gemma 4 is available on the Featherless platform, which offers seamless integration with existing OpenAI-compatible APIs, allowing users to switch with minimal code changes. The model features substantial improvements across various benchmarks, such as math reasoning and coding performance, and supports text, image, and audio inputs. It is available in four variants, including models optimized for edge and on-device use, and ships under the Apache 2.0 license, which encourages broader adoption by removing usage restrictions. Featherless provides a convenient and efficient infrastructure for using Gemma 4 without the need to manage complex GPU setups, making it an attractive option for production workloads focused on quality over raw throughput.