--- title: The video model race | daily.dev description: A roundup newsletter covering four newly launched AI video generation models: Seedance 2.5 (Bytedance), MiniMax H3, Flux 3 (Black Forest Labs), and Wan 3.0... canonical: https://daily.dev/posts/the-video-model-race-sn0dbk86l twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The video model race | daily.dev og:description: A roundup newsletter covering four newly launched AI video generation models: Seedance 2.5 (Bytedance), MiniMax H3, Flux 3 (Black Forest Labs), and Wan 3.0... og:url: https://daily.dev/posts/the-video-model-race-sn0dbk86l og:image: https://api.daily.dev/og/posts/SN0dbk86l.png og:image:alt: The video model race og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # The video model race **[Visually AI](https://daily.dev/sources/heatherbcooper)** · 8 min read · 0 upvotes · 0 comments ## Summary A roundup newsletter covering four newly launched AI video generation models: Seedance 2.5 (Bytedance), MiniMax H3, Flux 3 (Black Forest Labs), and Wan 3.0 (Alibaba). Each is broken down by duration limits, resolution, reference input capacity, native audio support, and per-second pricing, along with platform availability (Dreamina, CapCut, Fal, Krea, Magnific, etc.). The piece also shares example image and video prompts used with these models and closes with quick picks on InVideo Agent Two, the Freepik-to-Magnific rebrand, a new MiniMax Design desktop app, and LTX 2.5's open-weights world model. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the specs and pricing for Bytedance's Seedance 2.5 video model? Seedance 2.5 generates 4-30 second clips, extendable up to 30 additional seconds on Dreamina, at up to 1080p depending on the platform. It accepts 50 multi-modal references (30 image, 10 video, 10 audio), includes native audio, and costs about $0.47 per second at 720p. It's available on Dreamina, CapCut, Magnific, Fal, and Lovart. _Compare video model pricing and specs like these on daily.dev before picking a generator for a project._ ### How does MiniMax H3's resolution actually work at 4K? MiniMax H3 only natively renders at 480p and 768p; its 2K and 4K outputs are upscales of a 768p base rather than true native resolutions. It supports 5-15 second clips, 12 references (9 image, 3 video, 3 audio), native audio, a 7,000-character prompt cap, and costs about $0.13 per second at 2K on platforms like Hailuo and Krea. _Track native-versus-upscaled resolution claims across AI video tools on daily.dev before committing budget._ ### How does Wan 3.0's pricing compare to Seedance 2.5 for video generation? Wan 3.0 costs about $0.20 per second at 1080p, roughly a seventh of Seedance 2.5's cost for the same length. Wan 3.0 supports 2-30 second clips, 20 references (10 image, 5 video, 5 audio), plus decks, spreadsheets, PDFs and web pages as references, and native audio, though audio references can't be combined with other reference types. _Weigh cost and reference-limit tradeoffs between video models on daily.dev when choosing a generation tool._ --- Tags: [#genai](https://daily.dev/tags/genai), [#prompt-engineering](https://daily.dev/tags/prompt-engineering), [#ai-video-generation](https://daily.dev/tags/ai-video-generation) [View this post on daily.dev](https://daily.dev/posts/the-video-model-race-sn0dbk86l) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"The video model race","url":"https://daily.dev/posts/the-video-model-race-sn0dbk86l","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-video-model-race-sn0dbk86l"},"datePublished":"2026-08-31T13:08:36.613Z","dateModified":"2026-08-31T13:09:01.169Z","description":"A roundup newsletter covering four newly launched AI video generation models: Seedance 2.5 (Bytedance), MiniMax H3, Flux 3 (Black Forest Labs), and Wan 3.0...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/34d11c34e88c7363f2892e38c242f90c?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/34d11c34e88c7363f2892e38c242f90c?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Visually AI","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Visually AI","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/edab52f6b2824b42a36532e6659a3fc6","url":"https://daily.dev/sources/heatherbcooper"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-video-model-race-sn0dbk86l","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"genai,prompt-engineering,ai-video-generation","timeRequired":"PT8M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Visually AI","item":"https://daily.dev/sources/heatherbcooper"},{"@type":"ListItem","position":3,"name":"The video model race"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-video-model-race-sn0dbk86l#faq","mainEntity":[{"@type":"Question","name":"What are the specs and pricing for Bytedance's Seedance 2.5 video model?","acceptedAnswer":{"@type":"Answer","text":"Seedance 2.5 generates 4-30 second clips, extendable up to 30 additional seconds on Dreamina, at up to 1080p depending on the platform. It accepts 50 multi-modal references (30 image, 10 video, 10 audio), includes native audio, and costs about $0.47 per second at 720p. It's available on Dreamina, CapCut, Magnific, Fal, and Lovart. Compare video model pricing and specs like these on daily.dev before picking a generator for a project."}},{"@type":"Question","name":"How does MiniMax H3's resolution actually work at 4K?","acceptedAnswer":{"@type":"Answer","text":"MiniMax H3 only natively renders at 480p and 768p; its 2K and 4K outputs are upscales of a 768p base rather than true native resolutions. It supports 5-15 second clips, 12 references (9 image, 3 video, 3 audio), native audio, a 7,000-character prompt cap, and costs about $0.13 per second at 2K on platforms like Hailuo and Krea. Track native-versus-upscaled resolution claims across AI video tools on daily.dev before committing budget."}},{"@type":"Question","name":"How does Wan 3.0's pricing compare to Seedance 2.5 for video generation?","acceptedAnswer":{"@type":"Answer","text":"Wan 3.0 costs about $0.20 per second at 1080p, roughly a seventh of Seedance 2.5's cost for the same length. Wan 3.0 supports 2-30 second clips, 20 references (10 image, 5 video, 5 audio), plus decks, spreadsheets, PDFs and web pages as references, and native audio, though audio references can't be combined with other reference types. Weigh cost and reference-limit tradeoffs between video models on daily.dev when choosing a generation tool."}}]} ```