AI news ·
Pika Labs Releases Open Weight H3 Flex Video Model
Pika Labs released H3 Flex, an open-weight model that generates synced audio and video in six steps using multiple references.

Pika Labs has released an open-weight video model, H3 Flex, that generates synchronized audio and video. The model completes generation in six steps and supports multiple references as input.
The release makes media generation faster and more widely accessible. For professionals working across creative production, marketing, product development, events, science communication, and writing, this shift places greater competitive weight on how workflows are structured, how rights are managed, and how finished media is distributed.