HiDream Launches Flagship Image Generation Model HiDream-O1-Image-Pro
HiDream open-sources HiDream-O1-Image-Pro, based on the UiT native multimodal architecture, which unifies image pixels, text, and task conditions into a shared token space, discarding traditional VAE and separate encoders. It sets new SOTA on benchmarks such as GenEval and DPG, ranks 8th in the Artificial Analysis arena, and surpasses open-source models like FLUX.2[Dev].
HiDream open-sources HiDream-O1-Image-Pro, based on the UiT native multimodal architecture, which unifies image pixels, text, and task conditions into a shared token space, discarding traditional VAE and separate encoders. It sets new SOTA on benchmarks such as GenEval and DPG, ranks 8th in the Artificial Analysis arena, and surpasses open-source models like FLUX.2[Dev].
This page provides an editorial summary based on publicly available information. It is not a republished article. Use the source link below for the original report.