~2026-20928-12 来自 2026年4月5日 (日) 00:58

2026-04-05T00:58:32Z

新页面

{{Distinguish|Runaway}}{{Infobox company
| name = Runway AI, Inc.
| logo = Runway Logo.png
| logo_caption = 官方标志
| type = [[私营公司]]
| industry = [[人工智能]], [[机器学习]], [[软體开发]]
| founded = {{Start date and age|2018}}
| location = {{US}}[[纽约市]][[曼哈顿]]
| area_served =
| key_people = {{Unbulleted list|Cristóbal Valenzuela (CEO)| Anastasis Germanidis (CTO)| Alejandro Matamala (CDO)}}
| products = Gen-1, Gen-2, Gen-3 Alpha, Frames
| num_employees = 86
| website = {{URL|https://runwayml.com/}}
}}

'''Runway AI, Inc.'''（也称为'''Runway'''和'''RunwayML'''）是一家总部位于[[美国]]纽约市的科技公司。<ref name=":0">{{Cite news |last=Metz |first=Cade |date=2023-04-04 |title=Instant Videos Could Represent the Next Leap in A.I. Technology |language=en-US |work=The New York Times |url=https://www.nytimes.com/2023/04/04/technology/runway-ai-videos.html |access-date=2023-09-15 |issn=0362-4331}}</ref> 该公司主要致力于开发用于生成影片、图像和各种多媒体内容的产品和模型。它最著名的成就是开发了商业化的[[文本到影片生成模型]] Gen-1、Gen-2<ref name=":1">{{Cite news |date=2023-03-20 |title=Generative AI's Next Frontier Is Video |language=en |work=Bloomberg.com |url=https://www.bloomberg.com/news/articles/2023-03-20/generative-ai-s-next-frontier-is-video |access-date=2023-09-15}}</ref><ref name="Vincent">{{Cite web |last=Vincent |first=James |date=2023-03-20 |title=Text-to-video AI inches closer as startup Runway announces new model |url=https://www.theverge.com/2023/3/20/23648113/text-to-video-generative-ai-runway-ml-gen-2-model-access |access-date=2023-09-15 |website=The Verge |language=en-US}}</ref> 和 Gen-3 Alpha。<ref name=":0" />

Runway的工具和AI模型已被用于电影如《[[瞬息全宇宙|瞬息全宇宙]]》，<ref>{{Cite web |title=How director and editor Evan Halleck uses Runway for films, music videos, and commercials |url=https://runwayml.com/customers/how-director-and-editor-evan-halleck-uses-runway-for-films-music-videos-and-commercials/ |access-date=2024-03-01 |website=Runway |language=en-us}}</ref> 音乐影片如[[ASAP Rocky|A$AP Rocky]]、<ref>{{Cite web |title=Distorting Reality with Dan Streit |url=https://runwayml.com/customers/dan-streit/ |access-date=2024-03-01 |website=Runway |language=en-us}}</ref> [[Kanye West]]、<ref>{{Cite web |last=Cowen |first=Trace William |title=Kanye West Shares Unsettling Video for "Vultures (Havoc Version)" Ahead of Ty Dolla Sign Collab Album Release |url=https://www.complex.com/music/a/tracewilliamcowen/kanye-west-ty-dolla-sign-vultures-havoc-version-video |access-date=2024-03-01 |website=Complex |language=en-us}}</ref> [[Brockhampton (乐队)|Brockhampton]] 和 [[The Dandy Warhols]]，<ref>{{Citation |title=The Dandy Warhols - "I'd Like To Help You With Your Problem (feat. Slash)" - Official Music Video | date=14 February 2024 |url=https://www.youtube.com/watch?v=H3H2AVm1uD8 |access-date=2024-03-01 |language=en}}</ref> 以及编辑电视节目如《[[斯蒂芬·科尔伯特晚间秀|晚间秀]]》<ref>{{Cite web |title=How Runway took The Late Show edits from five hours to five minutes |url=https://runwayml.com/customers/late-night-show-cbs-uses-Runway/ |access-date=2024-03-01 |website=Runway |language=en-us}}</ref> 和《[[Top Gear (2002年电视剧)|Top Gear]]》。<ref>{{Cite web |title=How Drew Emery uses Runway to edit content for Top Gear America & Cooper's Bar |url=https://runwayml.com/customers/how-drew-emery-uses-runway-to-edit-content-for-top-gear-america-coopers-bar/ |access-date=2024-03-01 |website=Runway |language=en-us}}</ref>

== 历史 ==
该公司由智利人[[Cristóbal Valenzuela]]、<ref>{{Cite magazine |date=2023-09-07 |title=TIME100 AI 2023: Cristóbal Valenzuela |url=https://time.com/collection/time100-ai/6308278/cristobal-valenzuela/ |access-date=2023-09-15 |magazine=Time |language=en}}</ref> Alejandro Matamala 和希腊人Anastasis Germanidis于2018年创立，他们在[[纽约大学]]蒂施艺术学院相识。<ref>{{Cite web |last=Black |first=Julia |title='Not Everyone Is Trying to Build God': Runway CEO Cristóbal Valenzuela Tries to Dampen the Doomerism |url=https://www.theinformation.com/articles/not-everyone-is-trying-to-build-god-runway-ceo-crist-bal-valenzuela-tries-to-dampen-the-doomerism |access-date=2023-09-18 |website=The Information}}</ref> 公司于2018年筹集了200万美元，用于构建一个平台，以在多媒体应用中大规模部署机器学习模型。

2020年12月，Runway在[[A轮融资|A轮]]融资中筹集了850万美元。<ref>{{Cite web |date=2020-12-16 |title=RunwayML raises $8.5 million for its AI-powered media creation tools |url=https://venturebeat.com/business/runwayml-raises-8-5-million-for-its-ai-powered-media-creation-tools/ |access-date=2023-09-15 |website=VentureBeat |language=en-US}}</ref>

2021年12月，公司在B轮融资中筹集了3500万美元。<ref>{{Cite web |title=Runway raises $35M Series B {{!}} Runway Blog |url=https://runwayml.com/blog/runway-raises-35-series-b |access-date=2023-11-07 |website=Runway |language=en-us}}</ref>

2022年8月，公司与[[慕尼黑大学|路德维希马克西米利安大学]]的CompVis小组以及[[Stability AI]]的计算捐赠合作，发布了改进版的[[潜在扩散模型]] Stable Diffusion。<ref>{{Cite web |last=Cai |first=Kenrick |title=The AI Founder Taking Credit For Stable Diffusion's Success Has A History Of Exaggeration |url=https://www.forbes.com/sites/kenrickcai/2023/06/04/stable-diffusion-emad-mostaque-stability-ai-exaggeration/ |access-date=2023-09-15 |website=Forbes |language=en}}</ref><ref name="technologyreview.com">{{Cite web |title=The original startup behind Stable Diffusion has launched a generative AI for video |url=https://www.technologyreview.com/2023/02/06/1067897/runway-stable-diffusion-gen-1-generative-ai-for-video/ |access-date=2023-09-18 |website=MIT Technology Review |language=en}}</ref>

2022年12月21日，Runway在C轮融资中筹集了5000万美元，<ref>{{Cite web |last=Cai |first=Kenrick |title=Runway Raises
50
M
i
l
l
i
o
n
A
t
50MillionAt500 Million Valuation As Generative AI Craze Continues |url=https://www.forbes.com/sites/kenrickcai/2022/12/05/runway-ml-series-c-funding-500-million-valuation/ |access-date=2023-09-15 |website=Forbes |language=en}}</ref> 随后在2023年6月以15亿美元的估值进行了1.41亿美元的C轮扩展融资，<ref>{{Cite news |date=2023-06-29 |title=AI Video Startup Runway Raises 141M |url=https://techcrunch.com/2023/06/29/runway-a-startup-building-generative-ai-for-content-creators-raises-141m/ |access-date=2023-09-15 |website=TechCrunch |language=en-US}}</ref> 投资方包括[[Google]]、[[Nvidia]]和[[Salesforce]]。<ref>{{Cite web |title=Google Invests in AI Startup Runway to Wrest Cloud Business From AWS |url=https://www.theinformation.com/articles/google-invests-in-ai-startup-runway-to-wrest-cloud-business-from-aws |access-date=2023-09-15 |website=The Information}}</ref> 这些资金将用于构建用于电影和视频制作的多模态AI基础模型。<ref>{{Cite web |last=Black |first=Julia |title='Not Everyone Is Trying to Build God': Runway CEO Cristóbal Valenzuela Tries to Dampen the Doomerism |url=https://www.theinformation.com/articles/not-everyone-is-trying-to-build-god-runway-ceo-crist-bal-valenzuela-tries-to-dampen-the-doomerism |access-date=2023-09-15 |website=The Information}}</ref><ref>{{Cite web |title=Featured interview: Runway AI CEO Chris Valenzuela - First Move with Julia Chatterley - Podcast on CNN Audio |url=https://www.cnn.com/audio/podcasts/first-move-with-julia-chatterley/episodes/b9332b80-fc9b-483e-9ccf-b04c00f2a2c7 |access-date=2023-09-18 |website=CNN |language=en-US}}</ref>

2023年2月，Runway发布了Gen-1和Gen-2，这是首个商业化和公开可用的影片到影片和文本到影片生成基础模型，<ref name=":0"/><ref name=":1"/><ref name="Vincent"/> 用户可以通过简单的网页界面访问。

2023年6月，Runway被《[[时代杂志|时代]]》杂志评为全球100家最具影响力的公司之一。<ref>{{Cite magazine |date=2023-06-21 |title=TIME100 Most Influential Companies 2023: Runway |url=https://time.com/collection/time100-companies-2023/6285166/runway/ |access-date=2023-09-20 |magazine=Time |language=en}}</ref>

== 服务与技术 ==
Runway专注于影片、媒体和艺术的[[生成式人工智能|生成式AI]]。公司致力于开发专有的基础模型技术，供[[电影制作]]、[[后期制作]]、[[广告]]、[[编辑]]和[[视觉效果]]领域的专业人士使用。此外，Runway还提供一款面向消费者的[[iOS]]应用程序。<ref>{{Cite web |last=Vincent |first=James |date=2023-04-24 |title=Create generative AI video-to-video right from your phone with Runway's iOS app |url=https://www.theverge.com/2023/4/24/23695788/generative-ai-video-runway-mobile-app-ios |access-date=2024-03-01 |website=The Verge |language=en}}</ref>

Runway的产品可通过网页平台和[[API]]作为托管服务访问。

=== Stable Diffusion ===
[[Stable Diffusion]]是一个[[开源|开源]]的[[深度学习]]、[[文本到图像生成模型]]，于2022年发布，基于Runway和[[慕尼黑大学|慕尼黑路德维希-马克西米利安大学]]CompVis小组发表的论文《High-Resolution Image Synthesis with Latent Diffusion Models》。<ref>{{Cite web |date=2022-12-06 |title=Stable Diffusion developer Runway raises $50M to create AI multimedia tools |url=https://siliconangle.com/2022/12/06/stable-diffusion-developer-runway-raises-50m-create-ai-multimedia-tools/ |access-date=2023-09-18 |website=SiliconANGLE |language=en-US}}</ref><ref>{{Cite news |date=2023-08-08 |title=Stability AI Is Losing Executives, Engineers and Its Edge |language=en |work=Bloomberg.com |url=https://www.bloomberg.com/news/articles/2023-08-08/stability-ai-s-lead-threatened-by-departures-concerns-over-ceo |access-date=2023-09-18}}</ref><ref name="technologyreview.com"/> Stable Diffusion主要用于根据文本描述生成图像。

=== Gen-1 ===
Gen-1是一个影片到影片的[[生成式人工智能|生成式AI]]系统，通过将图像或文本提示的构图和风格应用于源影片的结构来合成新影片。该模型于2023年2月发布。Gen-1模型由Runway基于其研究论文《Structure and Content-Guided Video Synthesis with Diffusion Models》开发和训练。<ref>{{Cite web |title=Gen-1 by Runway |url=https://research.runwayml.com/gen1 |access-date=2023-09-18 |website=Runway |language=en-us}}</ref> Gen-1是影片创作的生成式人工智能的一个示例。

=== Gen-2 ===
Gen-2是一个多模态AI系统，可以根据文本、图像或影片片段生成影片。该模型是Gen-1的延续，并增加了根据文本生成影片的功能。Gen-2是首批商业化可用的[[文本到影片模型]]之一。<ref>{{Cite web |date=2023-03-22 |title=Text to Video Generative AI Is Finally Here and It's Weird as Hell |url=https://gizmodo.com/text-to-video-ai-art-generator-runway-modelscope-ai-1850249431 |access-date=2023-09-18 |website=Gizmodo |language=en}}</ref><ref>{{Cite web |last=Wiggers |first=Kyle |date=2023-06-09 |title=Runway's Gen-2 shows the limitations of today's text-to-video tech |url=https://techcrunch.com/2023/06/09/runways-gen-2-shows-the-limitations-of-todays-text-to-video-tech/ |access-date=2023-09-18 |website=TechCrunch |language=en-US}}</ref><ref>{{Cite web |author1=Avram Piltch |date=2023-06-08 |title=Runway's Powerful Gen-2 Text-to-Video Tool Now Available to Everyone for Free |url=https://www.tomshardware.com/news/runway-gen-2-text-to-video-available-to-all |access-date=2023-09-18 |website=Tom's Hardware |language=en}}</ref><ref>{{Cite web |last=Vincent |first=James |date=2023-03-20 |title=Text-to-video AI inches closer as startup Runway announces new model |url=https://www.theverge.com/2023/3/20/23648113/text-to-video-generative-ai-runway-ml-gen-2-model-access |access-date=2023-09-18 |website=The Verge |language=en-US}}</ref>

=== Gen-3 Alpha ===
Gen-3 Alpha是Runway即将推出的一系列模型中的第一个，这些模型在新的基础设施上进行了大规模多模态训练。它在保真度、一致性和运动方面比Gen-2有了显著改进，是构建通用世界模型的一步。<ref name=":1" />

Gen-3的训练数据来源于数千个[[YouTube]]影片，可能还包括[[盗版]]电影。一名前Runway员工向《[[404 Media]]》透露，公司曾通过[[youtube-dl]]和[[代理服务器]]下载影片，以避免被YouTube擋避。在测试中，《404 Media》发现输入某些[[YouTuber]]的名字会生成与其风格相符的影片。<ref>{{Cite news |last=Cole |first=Samantha |date=July 25, 2024 |title=AI Video Generator Runway Trained on Thousands of YouTube Videos Without Permission |url=https://www.404media.co/runway-ai-image-generator-training-data-youtube/ |access-date=July 25, 2024 |work=[[404 Media]]}}</ref>

=== AI电影节 ===
Runway每年在[[洛杉矶]]和[[纽约市]]举办AI电影节。<ref>{{Cite web |title=Runway AI Film Festival |url=https://aiff.runwayml.com/ |access-date=2023-11-18 |website=Runway |language=en-us}}</ref><ref>{{cite web |last=Melendez |first=Steven |date= |title=A new film festival will only show movies made using AI |url=https://www.fastcompany.com/90820457/a-new-film-festival-will-only-show-movies-made-using-ai}}</ref><ref>{{cite web |last=Kokalitcheva |first=Kia |title=1 big movie thing: an AI-generated film festival |url=https://www.axios.com/2023/03/26/ai-takes-on-film}}</ref>

== 参考文献 ==
{{reflist}}

== 外部链接 ==
*{{Official website}}

{{生成式人工智能}}

[[分类:纽约市公司]]
[[Category:2018年成立的美國公司]]
[[分类:人工智能公司]]

Runway - 版本历史

~2026-20928-12 来自 2026年4月5日 (日) 00:58