Baseten launches preview of server-side web search for open models
Baseten says Hosted Tools and Grounded Inference bring real-time search to open models running on its platform, claiming 15% lower latency than client-side execution.
TLDR
Baseten announced Hosted Tools and Grounded Inference in preview, saying they bring real-time web search server-side to open models through a single configuration. The company claims 15% lower latency than client-side execution, with no extra vendor key or orchestration required. Baseten says the preview launches with four web search partners: @ExaAILabs, @KeenableAI, @p0 and @youdotcom.
Combined views
32.4K
2 Sources, first seen ago