Its most significant model to date, Marengo 3.0 delivers human-like video understanding at enterprise-grade scale
SAN FRANCISCO, Dec. 1, 2025 /PRNewswire-PRWeb/ -- TwelveLabs, the leading video search and understanding company, announced today at AWS re:Invent general availability of its most sophisticated model yet, Marengo 3.0. The new release is a breakthrough video foundation model. It doesn't just watch video, it reads it, hears it, and picks up on the rhythm of a scene. The model can connect a moment of dialogue to a gesture three minutes later. It tracks objects, movement, emotion, and events through time. Simply, it is the world's most powerful video understanding model, and customers can access it today through Amazon Bedrock and TwelveLabs. To find out more about what makes Marengo 3.0 the world's most powerful video understanding model, please click here.
Built on TwelveLabs multimodal architecture, Marengo 3.0 uniquely treats video as a living, dynamic system, compressing audio, text, movement, visuals, and context into something that can be searched, navigated, and understood at scale. Marengo 3.0 comes production-ready and delivers immediate ROI. Based on extensive testing, the model offers 50% reduction in storage costs and 2x faster indexing performance among a slew of other benefits so that anyone with stores of video content can fully leverage all of their assets.
"Video represents 90% of digitized data, but that data has been largely unusable because it takes too long for humans to break down, and machines have been incapable of grasping and accounting for everything that happens in video," said Jae Lee, CEO and co-founder of TwelveLabs. 'Solving this problem has been our singular obsession. Now, Marengo 3.0 shatters the limits of what is possible. It is an incomparable solution for enterprises and developers."
Smarter, Faster, Leaner for True Video Understanding
The release of Marengo 3.0 positions TwelveLabs as the breakout leader in video intelligence infrastructure with capabilities no one else can match. Unlike competitors that rely on frame-by-frame analysis or separate image and audio models stitched together, Marengo 3.0 lets users see differently and understand everything in their video. This includes even the most complex, fast-moving clips.
Now, Marengo is even better at understanding sports, media & entertainment, and advertising video, as well as sensitive video types found across government and public security use cases. Marengo 3.0 delivers:
- Native Video Understanding: Marengo 3.0 was not adapted from image models. It offers understanding at the foundation model level.
- Temporal & Spatial Reasoning: The new model uniquely understands context across time and space.
- Sports Intelligence: In an industry-first, Marengo 3.0 offers team, player, jersey number, and action tracking to make identifying highlights faster and easier than ever before.
- Composed Multimodal Queries: To ensure users always find what they need, Marengo 3.0 enables them to combine image and text in a single query for more granular results.
- Production Economics: With 50% storage reduction costs and 2x faster indexing while creating the potential for new revenue streams, Marengo 3.0 is helping businesses save on cost while providing more opportunities for growth.
- Enterprise Ready: It's easy for even the largest organization to get started. Marengo 3.0 is available on Amazon Bedrock, enabling fast and secure deployment in their current AWS environment, as well as directly through TwelveLabs as a monthly service.
With its API-first design, Marengo 3.0 offers compact embeddings and four-hour video support– a 2x increase over Marengo 2.7. Additionally, it is multilingual across 36 languages.
"TwelveLabs' work in video understanding is transforming how entire industries manage their video capabilities, bringing unprecedented speed and efficiency to what has largely been a manual process," said Nishant Mehta, VP of AI Infrastructure at AWS. "We are excited to be the first cloud provider to offer Marengo 3.0 to our customers through Amazon Bedrock, following great adoption from TwelveLabs' previous Marengo and Pegasus models."
Marengo 3.0 is currently available through TwelveLabs or Amazon Bedrock, a fully managed service for building and scaling generative AI applications and agents. AWS is the first cloud service provider to offer access to Marengo 3.0.
About TwelveLabs
TwelveLabs is the world's most powerful video intelligence platform, enabling machines to see, hear, and reason about video like humans do. From semantic search to automated summaries and multimodal embeddings, TwelveLabs empowers developers and enterprises to unlock the full potential of video data across industries including media, advertising, government, security, and automotive. For more information, visit www.twelvelabs.io.
Media Contact
Amber Moore, Moore Communications, 1 5039439381, [email protected]
SOURCE TwelveLabs
Share this article