vidu q2 pro video Key Features

The vidu q2 pro video fast architecture is optimized for throughput. It can generate a 5-second cinematic 1080p clip in approximately 35 seconds. This makes vidu nearly three times faster than standard versions while maintaining professional-grade fidelity.

Yes, vidu is designed with advanced character anchoring. It maintains 98% visual identity consistency for human subjects across different shots. This allows for complex storytelling without the need for external LoRA training or specialized fine-tuning.

vidu supports generations up to 30 seconds per request. For longer narrative content, we recommend using the model's temporal consistency features to generate sequential clips and then stitching them together using traditional post-production tools.

Generation is priced at $0.15 per second of 1080p video. Reasoning tasks cost $2.50 per 1M input tokens and $10.00 per 1M output tokens. GPTProto users enjoy a 10% discount compared to direct retail pricing for enterprise-tier API keys.

We enforce a strict zero-data-retention policy for all vidu calls. Your input images, videos, and prompts are used solely for inference and are never used to train or fine-tune the base models, ensuring enterprise-grade privacy and security.

vidu natively understands fluid dynamics and object collisions. It scores 0.84 on the VBench Physical Logic set, meaning it produces realistic motion for things like splashing water or falling objects with 92% instruction-following accuracy.