News Summary:
On September 8, 2026, Witbe launched a family of Vision Language Models (VLMs) tailored for video test automation. These specialized models, deployed at NAB Show in Las Vegas, aim to run faster and cost less than general-purpose models, having been trained for the intricacies of set-top boxes and Smart TVs for more accurate and efficient testing. Previously, on September 7, Witbe announced its intention to showcase its AI-native Agentic AI platform at the IBC 2026 audiovisual trade show in Amsterdam. This platform, designed for automated testing and monitoring of video streams and services on real devices, will feature AI-driven test creation, hybrid AI, algorithmic scenarios, and the Smartgate AI analytics layer. Earlier, on September 6, Witbe released its internally-trained Vision Language Models (VLMs) for video test automation. These VLMs were developed and trained by Witbe, drawing on its experience testing and monitoring video services across devices, locations, and applications. This followed Witbe's initial unveiling of its own VLM family on September 4, expanding its agentic testing tools for video services. The company noted that these models are trained on various video service testing and monitoring cases, including on-screen keyboards, launcher grids, and frozen images on older smart TVs.
Subscribe for full access to Witbe's products in full detail 
