🚀 H3-World: Turning Language Understanding into World Control!<br>Instead of learning a new action interface from scratch, H3-World turns keyboard controls into language that MiniMax-H3 already understands, then grounds each instruction to the right moment in the generated video.</p>\n<p>✨ Highlights:</p>\n<ul>\n<li>Language-Native Control: Composes character and camera actions into textual instructions and injects them through MiniMax-H3’s pretrained text pathway.</li>\n<li>Temporally Grounded: Assigns one action prompt to each video latent interval, enabling precise control when actions change over time.</li>\n<li>Efficient &amp; Generalizable: Uses only 8,000 gameplay samples, 10,000 LoRA steps, and 0.199% trainable parameters to achieve controllable character and camera motion, including unseen action compositions and visual scenarios.</li>\n</ul>\n<p>👇 Dive in:<br> 📄 ArXiv: <a href=\"https://arxiv.org/abs/2609.01560\" rel=\"nofollow\">https://arxiv.org/abs/2609.01560</a><br> 💻 Code: <a href=\"https://github.com/Danzer1xxxxChan/H3-World\" rel=\"nofollow\">https://github.com/Danzer1xxxxChan/H3-World</a><br> 🏠 Project: <a href=\"https://github.com/Danzer1xxxxChan/H3-World\" rel=\"nofollow\">https://github.com/Danzer1xxxxChan/H3-World</a><br> 🤗 Model: <a href=\"https://huggingface.co/DANNY621/H3-World\">https://huggingface.co/DANNY621/H3-World</a></p>\n","updatedAt":"2026-09-02T02:19:06.294Z","author":{"_id":"668e740f1173ab43d9d9ed5e","avatarUrl":"/avatars/caa9b47c2a5f6d6d679759b8b234a0ab.svg","fullname":"Zeqing Wang","name":"INV-WZQ","type":"user","isPro":false,"isHf":false,"isHfAdmin":false,"isMod":false,"followerCount":11,"isUserFollowing":false}},"numEdits":0,"identifiedLanguage":{"language":"en","probability":0.753865122795105},"editors":["INV-WZQ"],"editorAvatarUrls":["/avatars/caa9b47c2a5f6d6d679759b8b234a0ab.svg"],"reactions":[],"isReport":false}}],"primaryEmailConfirmed":false,"paper":{"id":"2609.01560","authors":[{"_id":"6a97876cfe3c2f89286c392f","name":"Danze Chen","hidden":false},{"_id":"6a97876cfe3c2f89286c3930","name":"Zeqing Wang","hidden":false},{"_id":"6a97876cfe3c2f89286c3931","name":"Ziyue Lin","hidden":false},{"_id":"6a97876cfe3c2f89286c3932","name":"Xingyi Yang","hidden":false},{"_id":"6a97876cfe3c2f89286c3933","name":"Yeying Jin","hidden":false}],"mediaUrls":["https://cdn-uploads.huggingface.co/production/uploads/668e740f1173ab43d9d9ed5e/gnIBYQndPOGxRQ0GKIYzp.mp4"],"publishedAt":"2026-09-01T00:00:00.000Z","submittedOnDailyAt":"2026-09-02T00:00:00.000Z","title":"H3-World: Turning Language Understanding into World Control","submittedOnDailyBy":{"_id":"668e740f1173ab43d9d9ed5e","avatarUrl":"/avatars/caa9b47c2a5f6d6d679759b8b234a0ab.svg","isPro":false,"fullname":"Zeqing Wang","user":"INV-WZQ","type":"user","name":"INV-WZQ"},"summary":"We present H3-World, an efficient framework that turns the 33B MiniMax-H3 video generator into an interactive world model. Our key finding is that, as large video generators become more capable, language is emerging as a natural interface for control. MiniMax-H3, for example, already supports zero-shot control of character behavior and camera motion through natural-language instructions. Building on this, H3-World turns this coarse language interface into precise, temporally grounded world control, without introducing dedicated action modules. Specifically, we represent each action as a structured combination of character and camera instructions, and align them with the corresponding temporal video latents. To make the control temporally precise, we further introduce temporal attention routing, which restricts each instruction to its intended time interval and reduces control leakage across actions. Importantly, H3-World directly reuses the semantic representations learned during large-scale video pretraining and requires only lightweight adaptation. With only 8,000 gameplay samples, 10,000 LoRA optimization steps, and 0.199% trainable parameters, H3-World achieves effective character and camera control while preserving strong generation quality. It also generalizes to unseen scenarios. These results show that the control capabilities emerging in large video generators can be efficiently transformed into interactive world control.","upvotes":29,"discussionId":"6a97876cfe3c2f89286c3934","projectPage":"https://danzer1xxxxchan.github.io/H3-World/","githubRepo":"https://github.com/Danzer1xxxxChan/H3-World","githubRepoAddedBy":"user","githubStars":8},"canReadDatabase":false,"canManagePapers":false,"canSubmit":false,"hasHfLevelAccess":false,"upvoted":false,"upvoters":[{"_id":"668e740f1173ab43d9d9ed5e","avatarUrl":"/avatars/caa9b47c2a5f6d6d679759b8b234a0ab.svg","isPro":false,"fullname":"Zeqing Wang","user":"INV-WZQ","type":"user"},{"_id":"6729d1fed3ec5370cb035901","avatarUrl":"/avatars/50f7ce9c635148df76d1c63ebf3efa38.svg","isPro":false,"fullname":"1","user":"DANNY621","type":"user"},{"_id":"6a953f7cec1e4da705b805c7","avatarUrl":"/avatars/2cf6d1cf56ede343998fccba8430c0b6.svg","isPro":false,"fullname":"ZiyueLin","user":"RobinLinHK","type":"user"},{"_id":"634cfebc350bcee9bed20a4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/634cfebc350bcee9bed20a4d/fN47nN5rhw-HJaFLBZWQy.png","isPro":false,"fullname":"Xingyi Yang","user":"adamdad","type":"user"},{"_id":"6039478ab3ecf716b1a5fd4d","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6039478ab3ecf716b1a5fd4d/_Thy4E7taiSYBLKxEKJbT.jpeg","isPro":true,"fullname":"taesiri","user":"taesiri","type":"user"},{"_id":"65519eb532f278f503b3b2c3","avatarUrl":"/avatars/2e180f7b20189cd2d8a75e05c2913c5d.svg","isPro":false,"fullname":"lalala","user":"QuanjianSong","type":"user"},{"_id":"620783f24e28382272337ba4","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/620783f24e28382272337ba4/zkUveQPNiDfYjgGhuFErj.jpeg","isPro":false,"fullname":"GuoLiangTang","user":"Tommy930","type":"user"},{"_id":"63ca8e060609f1def7e6548a","avatarUrl":"/avatars/1da7947840cb87d5f77c0af9ee11f9c2.svg","isPro":false,"fullname":"Yi Jung","user":"YJ-142150","type":"user"},{"_id":"6624f53748e016b5ea587d40","avatarUrl":"/avatars/f8c16f45de0c3e32437f6e960a5b0959.svg","isPro":false,"fullname":"Shihua Zhang","user":"SuhZhang","type":"user"},{"_id":"677e31bb114aeff62d62bc11","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/JA85qWXtPO_mnK7_Bu2HJ.png","isPro":false,"fullname":"ZIJUN LIN","user":"onepiece1999","type":"user"},{"_id":"683725789cf5a24ae619ff81","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/no-auth/sE4DTVx1mCHbUf5uJjCcr.png","isPro":false,"fullname":"Kaiming Yang","user":"kaimingyang","type":"user"},{"_id":"6777819e38f9a731d4039535","avatarUrl":"https://cdn-avatars.huggingface.co/v1/production/uploads/6777819e38f9a731d4039535/QzpGJhc_9Fzp_yM1uB-Zi.jpeg","isPro":false,"fullname":"CaoZhijun","user":"aaroncaozj","type":"user"}],"acceptLanguages":["en"],"dailyPaperRank":0,"markdownContentUrl":"https://huggingface.co/buckets/huggingchat/papers-content/resolve/2609/2609.01560.md","query":{}}">

Get this paper in your agent:

Don't have the latest CLI?

Cite arxiv.org/abs/2609.01560 in a dataset README.md to link it from this page.

Add this paper to a collection to link it from this page.