Z.ai launched a new open-source image generation model, GLM-Image, that rivals proprietary systems like Google's Nano Banana Pro, also known as Gemini 3 Pro Image, in rendering complex text within visuals. The announcement, made this week, signals a growing trend towards open-source alternatives in the rapidly evolving field of AI image generation.

GLM-Image utilizes a hybrid architecture to achieve its proficiency in accurately rendering text-heavy infographics, making it suitable for enterprise applications such as collateral, training materials, onboarding documents, and stationary, according to VentureBeat. Google's Nano Banana Pro, part of the Gemini 3 AI model family released late last year, had previously been recognized for its speed, flexibility, and accuracy in this area.

While both Google's Gemini 3 and Anthropic's Claude Code have seen significant adoption and praise in 2026, these remain proprietary offerings. Z.ai's GLM-Image provides an open-source alternative, though VentureBeat noted that practical accuracy may still vary. The emergence of GLM-Image suggests that open-source development is keeping pace with proprietary advancements in AI image generation.