30 Oct 2026

Pika H3 Flex: No. 1 open-weight AI video model on DesignArena

Pika released Pika H3 Flex on 29 October 2026: a post-trained MiniMax H3 model that generates video with synchronized audio in six denoising steps instead of 50, with open weights (generator adapters and inference code) on Hugging Face and GitHub. Pika says it can make a 15-second 768p clip with audio in as little as seven seconds of transformer computation (a controlled benchmark, not end-to-end API latency), and that at the time of writing it ranked No. 1 among open-weights models and No. 7 overall on DesignArena's text-to-video leaderboard. It accepts up to 12 reference files mixing images, video and audio. The open weights run natively at 480p or 768p; the 1080p/2K/4K upscaler, five separate audio stems and prompt enhancement are Pika-hosted features, and the adapters are distributed under the MiniMax H3 Community License. Pika says physics, multi-scene consistency and complex reference control remain areas to improve.