Twelve Labs

Twelve Labs: Video AI Foundation Models

Twelve Labs is Video understanding foundation models — Marengo and Pegasus. Twelve Labs built foundation models specifically for video, where general-purpose LLMs underperform. The Marengo + Pegasus model family handles video search, classification, and generation natively — what most teams stitch together from Whisper + GPT + custom code.

Key Features

• Video understanding foundation models — Marengo and Pegasus
• Multimodal AI: video, audio, text reasoning in one model
• Search, classify, summarize, and reason about video content
• Used by media companies, security operations, and content platforms
• Radical Ventures and Index Ventures-backed
• Series B funded; $50M+ raised
• Targets the video-AI vertical where text/image LLMs fall short

Ideal Use Case

Media companies, content platforms, security operations, and any team working with video at scale that needs AI to understand video content — not just transcribe audio.

Why Use Twelve Labs

Twelve Labs built foundation models specifically for video, where general-purpose LLMs underperform. The Marengo + Pegasus model family handles video search, classification, and generation natively — what most teams stitch together from Whisper + GPT + custom code.

FAQ

Q: vs OpenAI Sora? A: Sora is generation; Twelve Labs is understanding. They're complementary, not competitive.

Q: Pricing? A: API + enterprise tiers; contact for current details.

Q: Languages supported? A: Multilingual including English, Spanish, French, German, Japanese, Chinese, Korean.

Q: Use cases? A: Content moderation, video search, sports analytics, media production, security ops.

tl;dr

Video understanding foundation models. Multimodal Marengo + Pegasus. Radical Ventures Series B.

Related

Looking for more options? Browse the AI/ML Models directory or read our best AI models listicle. Twelve Labs is also tracked on Crunchbase.