Google released four Gemma 4 models on April 2 — E2B, E4B, 26B MoE, and 31B Dense — built from the same research as Gemini 3 and licensed Apache 2.0 for unrestricted commercial use. The 31B Dense variant scores 85.2% on MMLU Pro, 89.2% on AIME 2026, and ranks third on Arena AI. All sizes support 256K-token context, native text/image/video/audio input, and 140+ languages.