Modular

The inference platform that replaced vLLM — pick it for yours

The inference platform that replaced vLLM — pick it for yours

Gemma 4 is live on Modular Cloud, day zero, with the fastest performance on both NVIDIA and AMD.Our MAX inference framework delivers 15% higher throughput vs. vLLM on B200, and we’re the only inference provider to ship Google DeepMind's Gemma 4 on a framework we built ourselves.Both flagship models available now:→ Gemma 4 31B: dense, 256K context, built for deep reasoning→ Gemma 4 26B A4B: MoE, 26B params, 4B active per forward passBoth natively multimodal: text, images, and video.Modular Cloud ...