MiniMax H3 视频模型登顶开源社区第一，定义视频模型领域“斩杀线”
Ranking
No observed public metrics; popularity remains neutral/archived.
Merged summary
TL;DR — MiniMax 开源多模态视频生成模型 H3，据报道登顶多项视频基准及 Hugging Face 热榜。其快速生态适配与低成本托管服务表明，视频模型竞争正从生成质量扩展到部署效率、开发者支持和商业化能力。
- H3 支持文生视频、图生视频、多参考生成、原生立体声音视频、多模态编辑和动作迁移。
- 据报道，H3 位列 Artificial Analysis 视频编辑基准和 Arena 图生视频榜首，并在开源后 24 小时内获得逾 100 家合作伙伴适配或集成。
- MetaSo 提供网页端、API 和 ComfyUI 接入，无需用户本地配置 GPU、CUDA 及相关依赖。
- 托管服务价格为 768p 每秒 0.09 元、2K 每秒 0.15 元；API 与网页端同价。报道中的 15 秒样例生成约需 3 分钟，而部分本地 10 秒视频估计需 20–40 分钟。
- H3 的生成质量、成本优势，以及芯片兼容、推理框架支持和应用部署速度，被视为推动企业采用与商业化的关键因素。
注： 雷峰网侧重榜单表现、生态适配和商业化，量子位更关注 MetaSo 的托管接入、功能、速度与定价。
Sources (2)
MiniMax H3 视频模型登顶开源社区第一，定义视频模型领域“斩杀线”
TL;DR - MiniMax open-sourced its H3 multimodal video-generation model, which reportedly topped multiple video benchmarks and Hugging Face’s trending rankings. Its rapid ecosystem adoption signals growing competition around deployment support and commercialization, not just model quality.
- H3 ranked first on Artificial Analysis’s video-editing benchmark and Arena’s image-to-video leaderboard.
- More than 100 partners reportedly adapted or integrated H3 within 24 hours.
- Ecosystem maturity increasingly depends on chip compatibility, inference-framework support, developer access, and application deployment speed.
- Analysts cited H3’s generation quality and cost advantages as potential drivers of enterprise adoption and commercialization.
等等，MiniMax H3不是刚发布吗？怎么就卷到几分钱的价格了……
TL;DR - MetaSo added MiniMax H3 video generation with web, API, and ComfyUI access, removing local deployment requirements. Pricing starts at RMB 0.09 per second for 768p and RMB 0.15 per second for 2K, lowering experimentation costs for creators.
- MiniMax H3 supports text-to-video, image-to-video, multi-reference generation, native stereo audio-video, multimodal editing, and action transfer.
- The hosted service avoids local GPU, CUDA, dependency, and configuration requirements.
- The article reports roughly three-minute generation for a 15-second example, versus an estimated 20–40 minutes locally for some 10-second outputs.
- Its official-compatible API uses the same pricing as the web service and can connect to ComfyUI workflows.