For the first time, an open video model sits at the same table as the closed frontier. That was the point of H3 - one model across tasks and modalitie...
@MiniMax_AI· 2026-07-31 12:29
7.6
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点视频生成已进入音画一体化阶段;开源权重让应用层企业摆脱闭源 API 依赖,显著降低垂直场景定制与推理成本。
Open weights. Shared progress. MiniMax H3 is moving fast. We built H3 for video generation with native stereo audio and multimodal reference control. ...
@MiniMax_AI· 2026-08-03 11:12
7.4
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Going live with @ComfyUI on Aug 7 at 10am PT to put H3 through its paces, in real time. → Open weights, native stereo audio, up to 2K/15s clips → Li...
@MiniMax_AI· 2026-08-04 08:18
6.4
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点这对 CEO 意味着应用层的多模态交互与内容生成门槛正加速降低,应关注其开源后在自有业务场景中的落地验证与成本优化空间。
👀 Stay tuned!ModelScope: ⌛️ Countdown to MiniMax H3's open source release 🔥 @MiniMax_AI 🔔 Follow for the drop. https://modelscope.cn/models/MiniMax/MiniMax-H3One model. Text, images, video, and audio in one unified context. Up to 15s of 2K video with native stereo sound.
@MiniMax_AI· 2026-08-01 09:52
7.0
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点通用大模型厂商正试图通过开源策略切入具身智能底层生态。对 CEO 而言,物理世界 AI 的基础模型壁垒正在降低,战略重心应转向高价值数据引擎构建与终端场景落地。
Robotics is one of the use cases we're most excited about for MiniMax H3. Open weights mean the embodied AI community can build data engines, world mo...
🔥 MiniMax H3 Max 发布
1 条最高 8.2
@MiniMax_AI· 2026-08-30 03:12
8.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
有些里程碑推动了排行榜的更新,极少数则改变了你对未来的看法。MiniMax H3 Max 兼具两者。我们现已将 MiniMax H3 作为开源权重模型发布……
对你意味着开源大模型选型库迎来高性能新选项。
事实MiniMax 宣布开源 MiniMax H3 Max 模型权重。
观点国内头部模型厂商加速开放高性能开源权重,正在挤压中游闭源 API 服务的溢价空间。CEO 应密切评估该模型在特定业务场景下的表现,探寻降低模型调用成本与防范单一供应商绑定风险的可能。
Some milestones move a leaderboard. A rare few change how you see the future. MiniMax H3 Max did both. We released MiniMax H3 as an open-weight model ...
🔥 MiniMax 视频模型开源
1 条最高 8.2
@MiniMax_AI· 2026-08-06 09:18
8.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
open weights soon𝔹𝕃𝕍ℂ𝕂𝕃!𝔾ℍ𝕋: I’m seeing way better videos coming from H3 than 2.5 on the timeline. And more of them. I think cost is a big part of it, plus the speed of inference, and custom audio uploads working as intended. My thesis holds still that at a certain point unless you are actually fixing
🔥 MiniMax Code CLI 发布
1 条最高 7.8
@MiniMax_AI· 2026-09-18 20:50
7.8
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
来自英伟达 Sol 团队针对 MiniMax-H3 的全新视频生成加速方案!生成速度快于播放速度,已完全开源,并可通过 Reactor API 调用...
对你意味着视频生成迈入实时化门槛。
事实MiniMax-H3 经英伟达加速后推理速度超过播放速度且已开源。
观点视频生成突破实时瓶颈后,企业构建流式视频应用的成本与时延正显著下降。
New video generation acceleration for MiniMax-H3 from the NVIDIA Sol team! Faster than playback, fully open-sourced, and available via the Reactor API...
🔥 MiniMax Fast H3 发布
1 条最高 7.6
@MiniMax_AI· 2026-09-03 03:17
7.6
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
这就是开源智能的意义所在。基于 vLLM-Omni 与 @haoailab 的 FastH3 构建,并得到了 NVIDIA AI 的硬件支持——全部完全公开。实时生成……
言外之意,模型厂商正通过整合开源推理生态打通端到端实时生成链条。
事实MiniMax 宣布公开基于 vLLM-Omni 与 FastH3 的实时生成架构。
观点这意味着 CEO 在开发实时 AI 交互应用时,可直接利用成熟基础设施降低研发与推理成本。
This is what open-source intelligence is for. Built on vLLM-Omni and @haoailab's FastH3, with @NVIDIAAI hardware support - all public. Real-time gener...
🔥 MiniMax H3 登顶 Video Edit Arena
1 条最高 7.6
@MiniMax_AI· 2026-08-14 00:25
7.6
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Now available in @LumaLabsAI Agents: H3 generating up to 15s of 2K video with native stereo sound. More places to direct, create, and bring ideas to l...
🔥 Maestro v1.5.5 发布
1 条最高 7.2
@MiniMax_AI· 2026-08-05 04:56
7.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Open weights are step one. Making them easy to serve is step two. H3 now has Day 0 support in @vllm_project's vLLM-Omni, complete with an OpenAI-compa...
🔥 MiniMax H3 上线
1 条最高 7.2
@MiniMax_AI· 2026-08-01 00:34
7.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
From a single clip to finished content with sound, characters, and creative intent intact. MiniMax H3 is now live on @LeonardoAi. Open weights coming ...
🔥 MiniMax H3 上线与开源计划
1 条最高 7.2
@MiniMax_AI· 2026-08-01 00:04
7.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
H3 模型现在只需一次 generateVideo() 调用即可使用 ▲ 已在 Vercel AI Gateway 上线。非常感谢 @vercel_dev 和 @rauchg 的 Day 0 支持。开源权重即将发布...
事实MiniMax 将 H3 视频生成模型直接接入 Vercel 开发生态并规划开源。
观点视频大模型正在从闭源 API 竞争走向基础设施分发与生态化落地。对 CEO 而言,多媒体 AI 功能的集成门槛和成本大幅下降,应评估在自身应用中引入统一 Gateway 架构以快速测试视频能力。
H3 is now just one generateVideo() away ▲ Live on Vercel AI Gateway. Huge thanks to @vercel_dev and @rauchg for the Day 0 support Open weights coming...
🔥 MiniMax 2K原生立体声模型发布
1 条最高 7.2
@MiniMax_AI· 2026-07-31 13:14
7.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
我们很荣幸能作为 AI 合作伙伴参与新推出的 @Singtel AI Pass,支持新加坡的 SkillsFuture AI 订阅计划。
对你意味着
事实MiniMax 绑定海外电信巨头切入政府技能补贴体系。
观点海外直接获客成本高企下,借力本地运营商套餐与政府资金做 B2B2C 捆绑,是 AI 产品快速渗透区域市场的关键策略。
We're proud to contribute as an AI partner to the newly launched @Singtel AI Pass, supporting Singapore's SkillsFuture AI Subscription initiative. �...
🔥 MiniMax H3 Mac 推理引擎发布
1 条最高 6.4
@MiniMax_AI· 2026-08-11 06:17
6.4
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Across infrastructure, creative platforms, and open communities, our partners share a belief in what open models can unlock. Grateful to @fal, @pika_l...
🔥 MiniMax GMI Cloud 限时免费活动
1 条最高 5.6
@MiniMax_AI· 2026-08-25 02:01
5.6
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点这表明国产模型厂商正加速融入全球 AI 创作者生态,CEO 可关注其模型在真实设计与渲染工作流中的集成表现。
Waitlist is open for tonight for our H3 x @magnific showcase but that doesn’t mean you’re out. Register now for a chance to get in: https://luma.com...
🔥 MiniMax H3 挑战赛
1 条最高 4.6
@MiniMax_AI· 2026-08-21 05:53
4.6
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点头部模型厂商正加速推动生态绑定战略。对 CEO 而言,选择技术迭代稳定且重视开发者生态的基座厂商,有助于在模型能力演进中获取持续的技术红利与合作资源。
The road to AGI was never meant to be traveled alone. From Hailuo 01 to H3, we've been incredibly lucky to build alongside so many friends, partners, ...
@MiniMax_AI· 2026-09-16 02:15
6.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Have you ever felt scared by how powerful your model is? This really makes me wonder: are we living in a simulation? The whole world may just be a sta...
@MiniMax_AI· 2026-08-18 09:47
3.8
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
这太酷了!像我们的 MiniMax H3 这样的 SOTA 视频生成模型能够非常好地理解行走人物的运动,尽管有些人反馈……
事实MiniMax 展示了其 H3 模型在人体运动理解上的表现。
观点视频生成正由画质渲染转向物理运动规律建模。对于 AI 应用团队,需密切关注国产视频模型在复杂动作生成上的实际落地表现。
This is super cool! Sota video generation models like our MiniMax H3 understand the movement of a walking character so well, while some people report ...
@MiniMax_AI· 2026-08-12 07:53
6.4
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点对应用层 CEO 而言,依赖单一端到端生成模型往往高昂且难以控场,采用搭积木式的复合工作流才能在商业落地中兼顾极致成本与精确可控性。
Let H3 cook!! 🧑🍳🔥GMI Cloud: you don't need Seedance 2.5, here's cheaper and more efficient workflowwe made this entire game shooting scene in one shot for $1.97 with DeepSeek V4 Flash 0731 + MiniMax H3we use DeepSeek to code the raw Three.js scene first: exact camera movement, character motion, timing,
@MiniMax_AI· 2026-08-15 17:29
6.0
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
The real breakthrough in the NVIDIA SANA team’s Sol Engine work on MiniMax H3! By splitting generation into a 4-step low-res H3 draft and a 3-step L...
@MiniMax_AI· 2026-08-22 04:07
6.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Four days after we opened the weights, the community built what's usually a lab deliverable: a distillation LoRA that cuts sampling from 20 steps to 4...
@MiniMax_AI· 2026-09-02 06:26
6.4
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
观点对 AI 应用层 CEO 而言,多模态视频领域的开放生态正在成熟,基于开源或开放 API 构建自控多模态工作流的战略可行性显著增加。
We would never see this level of creation everywhere if SOTA video models stayed behind closed doors. Every single day, we're blown away by what you'r...
@MiniMax_AI· 2026-08-06 02:10
6.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
Three paths to faster video attention: compute the same interactions more efficiently, compute fewer in full, or change how information is mixed. Here...
📘 大模型推理优化
1 条最高 6.2
@MiniMax_AI· 2026-08-20 05:27
6.2
@MiniMax_AI
MiniMax · 88K followers
MiniMax 官方账号,中国六小虎之一,旗下 Hailuo 视频与 abab 文本模型在 C 端出圈;官方推文主要释放产品更新与 API 动态,是追踪 MiniMax 商业化节奏的窗口
RT RonnyHe: The previous article covered the harness behind MiniMax Code. This time I want to take on a harder question: after changing the harness, h...
RT Tom Shapland: I’m thrilled to launch Faxie! Faxie is an app for generating videos inside Slack. Our vision is to help hybrid teams feel more conne...
RT GMI Cloud: For our MiniMax Week challenge with @MiniMax_AI, we asked builders to create with MiniMax models, from LLMs to multimodal, across three ...
RT GMI Cloud: Tomorrow, we’re building with H3 alongside the @MiniMax_AI team: from prompt structure to a finished video generation on GMI Cloud. We...
RT RyanLee: This is my favorite kind of open-source result. We released H3 at 28 steps. Then the community started optimizing it. Now, after 11,850 vo...
RT Minds by Animoca Brands: "Most users are going to have a relationship with the agent, so we're going to give them best in class and best in price."...
RT fal: Introducing fal's H3 Max Director, the first natively continuous real-time frontier video model for action-controlled long-form generation. Un...
We’re proud to see MiniMax-M3 powering HUMAIN-M3. Built on the M3 foundation and further trained on more than 1 trillion Arabic tokens, HUMAIN-M3 bri...
RT GMI Cloud: Join us this Thursday with @MiniMax_AI for a live session on M3 for agents, coding + reasoning, plus a live build using GMI API key. Sep...
RT GMI Cloud: 5 more days of FREE MiniMax M3, M2.7, Music 3.0, and Speech 2.8 on GMI Cloud also, join the MiniMaxathon to get $500 cash, 3 months of M...
RT ComfyUI: See the best @minimax_AI H3 workflows in action as we announce winners of our #ComfyH3 Sync Sound challenge on a special livestream with g...
RT Miora Design: Bestiary: Miora × MiniMax H3 Creative Contest is now open. Every creature comes from an encounter no one wrote down. A legend, a dre...
RT Leanna Ren: ps It feels so damn good to be creating and driving new product paradigms, especially with good friends and partners. also the power of...
RT Leanna Ren: This is exactly why we open source! You build the model, put it in people’s hands, and watch them take it somewhere magically unexpect...
A great deep dive from the @fal team on how they built H3 Max, combining post-training with a co-designed inference stack to improve prompt adherence,...
Congrats to the fal team on MinMax-H3 Max - fantastic work! fal optimized the model for stronger real-world performance while co-designing the inferen...
RT Leanna Ren: MiniMax H3-Max combines SOTA video quality with a step-change in generation speed, making high-quality video practical for many more re...
The MiniMax H3 ecosystem is growing faster than ever! ☺️ Check out Awesome MiniMax H3 Integrations - an index tracking everything built around H3, f...
RT ComfyUI: Livestream alert! Local MCP: Run ComfyUI with your Agent & Hardware on Weds. 8/26 at 10am PT. Join us for a live conversation on how to ge...
RT Magnific: This video was almost faster to make than to watch Unlimited MiniMax H3 Max for the next 3 days, exclusively in Magnific → Generate as m...
RT Leanna Ren: 2 yrs ago when we first launched @Hailuo_AI, @runwayml has been setting an incredibly high bar for content, aesthetics, and technical q...
RT OpenArt: ⚡ OpenArt Unlimited Week is here! Upgrade to an eligible Pro or Wonder plan by August 23 to unlock 7 days of unlimited generations on the...
RT ComfyUI: Seedance 2.5 tends to pick up more natural movement and follow 3D blockouts more closely—but sometimes the results are nearly 1:1 with Mi...
RT Replicate: H3 from @Hailuo_AI is live on Replicate! The latest from Minimax, particularly well-tuned for reference-based creation and video editing...
RT Leanna Ren: This was from last summer when our cofounder, sales lead, and I (all females) were on a roadshow, I crashed on different sofas for two ...
RT Leanna Ren: While I just landed in SF and open sourced our music model this morning, there are already so many music creations out there! Happy ope...
RT Modular: Building frontier models involves a hundred decisions rarely discussed publicly: when to stop scaling, what data to keep, how to different...
RT Wayne Sutton: Packed at @convex hq tonight for the AiOS Meetup - San Francisco Edition meetup hosted by @ronaldmannak @RayFernando1337 @VictorSuOrt...
RT Leanna Ren: Half a year ago we open-sourced our Visual Tokenizer Pre-training (VTP) work, the first to validate a scaling law for visual tokenizers...
RT SambaNova: SN50 is REALLY FAST. And you don't have to just take our word for it. ⚡️ @SemiAnalysis independently benchmarked SN50 running @MiniMax...
RT RyanLee: None of this was us. Day 1 (and the first 48 hours) belonged to the open-source community: Generate with it @ComfyUI — native support + o...
RT RyanLee: We just dropped SOTA open weights for our MiniMax‑H3 video model, and so much happened in 48 hours. The community ran H3 on untested hard...
RT Victor Su-Ortiz: H3 can be licensed for deployment in the US, EU, UK, and South Korea through MiniMax’s formal authorization process. Apply by ema...
RT MTS: SITUATION DETECTED: MiniMax-H3 is now the #1 open model in Video Arena across both Text-to-Video and Image-to-Video, and is now tied for #3 Ov...
RT ComfyUI: MiniMax H3 open weights are out, with the nodes already in ComfyUI. One Blender blockout, run against Seedance 2.0 at a fraction of the co...
RT Leanna Ren: People who know me well know that both @Hailuo_AI and @MiniMax_AI have been my treasures since day one (especially Hailuo)! I’ve been ...
RT Magnific: One Odysseus in nine scenes We locked one character, and made the movie with MiniMax H3. Powered by @MiniMax_AI Same face in every scene....
RT Factory: Join us August 6th for an evening of Open Source with @MiniMax_AI, @Kimi_Moonshot, @baseten, @modal, and other builders. We have one exclu...
RT PixVerse: MiniMax H3 is now live on PixVerse! @Hailuo_AI 2K video, native audio, and multimodal references in one workflow Members can get 50% off ...
RT Pollo AI: MiniMax H3 is live on Pollo, with a 50% launch offer. Omni-reference input. Production-ready output. Up to 15-second videos in 2K, with n...
RT Magnific: MiniMax H3 syncs voice to lips natively, per character, in one generation This 4-shot scene took one prompt and kept everything consisten...
RT RyanLee: Great to see @FireworksAI_HQ continuous optimizations on MiniMax Sparse Attention (MSA). By refining the attention kernel’s load and stor...
RT Starchild: MiniMax M3 is live on Starchild. It's built for long tasks: holding a lot of context, working across many steps, calling tools and reaso...
RT SambaNova: Most AI apps resend the same system prompt or reference doc on every call, and pay full price to reprocess it every time. SambaCloud's n...
RT Kilo: Across a full year of AI usage on @kilocode (July 2025 to July 2026), one lab used more tokens than anyone else, by a wide margin. @MiniMax_A...