MiniMax launches open-weight H3 Image model
MiniMax released H3 Image for image generation and editing, with model weights available on launch day, API access, and support through MiniMax's developer and Hailuo platforms. 爱范儿 reported that its hands-on tests generated 4096×4096 images in 10–20 seconds and highlighted support for up to 16 reference images, while MiniMax's claim of state-of-the-art performance among open-weight image models was not independently verified.
H3 Image gives developers open-weight access to a high-resolution image generator that can also be connected to coding-agent workflows for iterative visual creation.
今年 8 月，MiniMax 开放了 H3 视频生成模型的权重，这是有史以来第一个达到 SOTA 水平的开源视频模型。
此后几个月，社区基于它玩出的花样远超官方演示，也催生了 H3 max 等一批衍生模型，实实在在推着 AI 创意行业往前走了一步。
现在，MiniMax 连图像生成模型也不打算藏着了。就在刚刚，MiniMax H3 Image 正式发布，重点是图像生成与编辑，目标是开源图像模型中的 SOTA。
此前的 H3 已经通过统一的全模态预训练，具备了图像、视频、音频等多种内容的理解与生成能力，其中也包括文生图和图像编辑。
Some source content is omitted. Continue reading on the original site.
#MiniMax#MiniMax H3 Image#MiniMax H3#MiniMax Code#MiniMax Hailuo