SponsoredAI 会员在线充值
Visit sponsor Chatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心Loading…
SponsoredChatgpt/Claude/Grok 会员24h在线自助充值 安全快速 稳定保障
奥创(Ultra)商城会员充值中心
Sponsored本页公开展示实时样本、模型覆盖与响应时间。先用 5 元按量跑自己的 Claude Code/Codex 任务，成功后再决定是否长期使用。
ACS Gateway
Sponsored【腾讯云】WorkBuddy 全场景 AI 办公工作台，团队旗舰版低至198元/月
腾讯云
Sponsored人人可用的AI应用创作平台 自然语言交互，全栈应用搭建，伟大创意快速实现;个人版新用户注册送5000积分
阿里云MiniMax says Morgan Suo will discuss how the company’s H3 model is advancing AI video generation at ModCon in San Francisco on August 18.
Morgan Suo, MiniMax’s head of US business development, is scheduled to speak at ModCon 2026 about H3 and the role of open-weight, omni-context generation in AI video workflows. The event is set for August 18 in San Francisco.
Available product documentation describes H3 as a multimodal video model with text-to-video, keyframe-based generation, and an Omni Reference workflow. It can use image, video, and audio references to guide subjects, motion, camera work, sound, and visual atmosphere, with outputs ranging from 5 to 15 seconds and native stereo audio support.
Source review does not guarantee permanent accuracy. Consider the original source, observation time, and later corrections.
ModCon 2026 is August 18th at the Grand Hyatt in San Francisco, and is our flagship developer conference for the people building, running, and
H3 can replace, add, or remove characters and objects; change backgrounds, lighting, atmosphere, and visual effects; and revise dialogue, human voice, or voice identity. Its precise instruction-following is designed to apply targeted changes while keeping unedited content relatively stable. H3 generates videos from 5 to 15 seconds at 24 FPS, and every result includes native stereo sound. It supports common ratios including 21:9, 16:9, 4:3, 1:1, 3:4, and 9:16, with 1440p output available. The product documentation lists 768p as a later rollout option. [...] H3 supports Text-to-Video, first-frame and first-and-last-frame generation, and an Omni Reference workflow. Omni Reference accepts up to 9 images, 3 video clips, and 3 audio clips, with a combined limit of 12 files. Audio references must be used together with an image or video. You can reference character appearance, actions, camera movement, composition, voice, visual atmosphere, narrative rhythm, sound design, and editing style.
Morgan Suo, Head of US Business Development, joins ModCon '26 to share what open-weight omni-context generation enables for AI video.
Join us at ModCon on August 18, 2026 in San Francisco for a full day of AI infrastructure launches, technical deep dives, hands-on demos,
Watch the full ModCon 2026 keynote, recorded live at the Grand Hyatt in San Francisco on August 18, 2026.
4 months ago, we could see things are changing every month, every few weeks. Now for every single episode, things are changing dramatically. Dramatically, yeah. Like the difference is not uh you know small percentage. from a minute to 50 seconds, a minute to 6 seconds. And I think for people who are watching the show, this is great that you guys are keeping yourself up to date even if you are not trying. Even if you are not trying, if you are not going to the file, knowing that what is happening and then applying it to different things. This is happening to the video today. It happens to image. It happens to code. It happens to whatever industry you are in. So, things are changing really fast and it's really important to understand how to adapt, how to change your business process. It [...] But the quality of the videos has actually improved in terms of prompt adherence compared to the original model. That takes much longer to generate the videos. Look at that third row there. Average