Baseten Expands Model Library With Early Access to Gemma 4 Multimodal AI

According to a recent LinkedIn post from Baseten, the company’s platform now provides day-zero access to Google’s Gemma 4 family of multimodal AI models through its managed model library. The post indicates these models support text and image inputs with text outputs, emphasizing capabilities such as advanced reasoning, coding and function calling, OCR for document understanding, and long context windows up to 256K tokens.

The company’s LinkedIn post also highlights Gemma 4’s architectural innovations, including alternative attention mechanisms, Proportional RoPE, Per-Layer Embeddings, KV-cache sharing, native aspect ratio handling for vision, and a smaller frame window for audio, all framed as targeting greater efficiency and scalability. For investors, this suggests Baseten is positioning itself as an early mover in hosting cutting-edge open models, which could enhance its appeal to AI-driven enterprises, deepen usage of its infrastructure, and potentially improve customer retention and monetization as demand for high-performance multimodal workloads grows.