Navigating the rapidly expanding landscape of artificial intelligence has become increasingly challenging for everyday users, as high-tech firms continuously roll out new models. To address this complexity, Vancouver-born startup Vresk has launched a web-based workspace designed to help users directly compare the outputs of different AI models side by side.
Developed by UBC Sauder School of Business graduate Josh Hikida and crafted by Baran Labs, Vresk allows users to pose a single question or prompt to up to three open AI models simultaneously. The platform supports text and building capabilities, enabling users to evaluate models like Kimi K3, DeepSeek V4 Pro, Qwen 3.5, and GLM 5.3 while maintaining files and memory within a unified workspace.
According to Hikida, the tool is intended to help users determine which model best fits their specific workflow by observing exact answers side by side. Vresk is available to users for free, alongside a "Plus" tier priced at US$20 per month.