xAI deprecating Grok API Models – Grok 2, 3 Mini and Grok 2 Vision

xAI has notified developers that it is deprecating three models in its Grok API: grok-2-image-1212, grok-3-mini, and grok-2-vision-1212. These models will no longer be supported after February 28, 2026, with xAI encouraging migration to newer alternatives for improved performance and capabilities.

The grok-2-image-1212, launched in December 2024, is an image generation model designed to produce multiple high-quality images from text prompts. It formed part of xAI’s expansion of multimodal features in the API, enabling developers to integrate advanced visual creation into applications.

Released on December 12, 2024, alongside related updates, the grok-2-vision-1212 is a multimodal vision model that supports image inputs alongside text. According to official announcements, it offers enhancements in accuracy, instruction-following, and multilingual processing, making it suitable for tasks involving visual understanding and analysis.

The grok-3-mini is a lightweight variant associated with the Grok-3 series, which debuted in February 2025. Grok-3 itself marked a significant advancement in reasoning capabilities, positioning xAI at the forefront of AI development.

The mini version provided efficient performance for resource-constrained applications while retaining strong reasoning and response quality. xAI advises transitioning image generation workflows to the newer grok-imagine-image model, which delivers substantially improved output quality.

For video-related tasks, the recently introduced grok-imagine-video model supports advanced creation and editing features, including high-resolution outputs and better prompt adherence.

This deprecation reflects xAI’s focus on evolving its API toward more capable models, as seen in recent releases like Grok Imagine 1.0. Developers are directed to the xAI console and documentation for migration guidance and to monitor for deprecated tags on model listings.

(source)