Pegasus 1.6
#17 en Modèles multimodauxTwelvelabs · v1.6 · depuis 6. Oktober 2026 · 3× · vu le 13 oct. 2026
28
Momentum
Pegasus 1.6 is a video-to-text model from TwelveLabs (Seoul/San Francisco), released on October 6, 2026. It is the company's first model built for egocentric (first-person) video and also supports native image analysis. Access is via the TwelveLabs API; according to one source, batch analysis is still handled by Pegasus 1.5.
Historique du momentum
20.07.18.10.
Fonctionnalités
| Context Window (Tokens) | 261,120 tokens (videos up to ~2 hours) |
| Multimodality | Video and images (text output); native image understanding for on-screen text, logos |
| Platform | TwelveLabs API |
| Price per 1M Tokens | Output: $7.50 per 1M tokens; input image: $3 per 1M tokens; video: $1.75 per video hour |
| Release Date | October 6, 2026 |