Imagine Image 2.0 Ranks Second Globally on Text-to-Image Arena Benchmark
xAI has released Imagine Image 2.0, which has achieved the #2 rank globally for both text-to-image generation and image editing on the Arena benchmark. The model is currently available as the "Quality Mode" on Grok's consumer applications. This release positions xAI's image generation capabilities directly behind OpenAI's GPT-Image-2, intensifying the competition among top AI labs. It highlights the rapid advancement of high-fidelity image generation and native image editing tools integrated directly into consumer AI assistants. Imagine Image 2.0 features advanced editing capabilities such as Magic Wand and Multi-Ref Editing, and is trained for high fidelity across photography, design, and illustration. While currently accessible via Grok apps, developer API access is expected to be released soon.
## BACKGROUND
The Text-to-Image Arena is a crowdsourced benchmark platform where users compare side-by-side image outputs from different AI models to rank them based on human preference. xAI, founded by Elon Musk, develops the Grok AI assistant, which integrates these generative models to compete with rivals like OpenAI and Google.