Huawei Hong Kong AI Lab open-sources the 100B-parameter image model Boogu-Image-0.1

According to Beating Monitor, the Boogu team, led by Huawei Hong Kong’s AI Lab and involving researchers from multiple universities, has open-sourced Boogu-Image-0.1—a 10-billion-parameter image generation and editing model comprising Base, Turbo, Edit, and Edit-Turbo variants, supporting text-to-image generation in both Chinese and English. The base model was trained from scratch using only 208.62 million deduplicated images, with an estimated training cost of approximately $400,000. The paper states that Boogu-Image-0.1 achieves top-tier performance among open-source models across multiple standard benchmarks, with some results approaching those of leading proprietary models. The Base and Edit variants support up to 2K resolution, while Turbo generates 1K images in just 3 to 4 steps. The system first understands and rewrites user prompts, then selects the appropriate model based on task complexity to minimize unnecessary inference costs. Model weights, inference code, and some training methods are released under the Apache 2.0 license.