Video and Image Generation: Sora, Veo, Kling, Flux — 2026-10-10
Vidu Q4 Preview launched with 4K output and multi-reference capabilities, while a US appeals court ruled against fair use for AI training data. Meanwhile, Black Forest Labs updated its FLUX.2 models, and industry analysts highlighted the shift toward service-based revenue in AI video.
Video and Image Generation: Sora, Veo, Kling, Flux — 2026-10-10
Vidu Q4 Preview launches with 4K and multi-reference support
On October 10, 2026, Beijing-based Shengshu Technology released the preview version of its next-generation video generation model, Vidu Q4. The model supports up to 15 reference images and 3 reference audio tracks, enabling complex character and scene consistency. It offers output resolutions up to 2K and 4K, positioning it as a high-fidelity competitor to models like Kling and Veo.
US Appeals Court rules AI training data is not fair use
In a significant legal development reported on October 6, 2026, a US appeals court upheld a ruling that using copyrighted material to train AI models does not constitute "fair use." The case involved Ross Intelligence, an AI legal research company, but the precedent has broad implications for generative AI firms that rely on scraped web data. This decision intensifies legal pressure on video and image generators regarding their training datasets.
Black Forest Labs updates FLUX.2 models
Black Forest Labs released updates to its FLUX.2 [pro] and FLUX.2 [flex] image generation models. The release focuses on production-grade image generation at scale, with pricing starting at $0.03 per megapixel. This update aims to provide high-quality generation at affordable prices, competing directly with other leading image APIs.
Industry report: Revenue comes from delivery, not just models
A report published by Phoenix News (ifeng.com) on October 4, 2026, analyzed 26 AI video products. It found that while many companies integrate third-party models, the top earners—such as Kling, Jimeng, and Hailuo AI—are those with self-developed video models. However, the report noted that revenue is increasingly driven by engineering and delivery services rather than raw model access alone.
Local view
Chinese media outlets are closely tracking the domestic competition between Kuaishou's Kling, ByteDance's Jimeng, and other local players. A recent analysis by Phoenix News highlights that while Kling and Hailuo maintain strong model capabilities, the market is consolidating around platforms that offer integrated creative workflows. The release of Vidu Q4 is seen as a direct response to the need for higher resolution and better control in professional production pipelines.
Context & numbers
- Resolution Standards: Leading models like Vidu Q4, Kling 3.0, and Veo 3.1 now support 4K output, though pricing tiers often restrict native 4K to premium subscriptions or per-second billing.
- Pricing Benchmarks: Recent comparisons indicate that image generation via FLUX.2 starts at $0.03/MP, while video generation costs vary significantly by length and resolution, with some models offering "per-second" pricing structures that can exceed $0.20/sec for high-end features.
- Arena Rankings: Artificial Analysis leaderboards continue to track models like Gemini Omni Flash and Wan 3.0, with Elo scores fluctuating based on user preference for motion coherence and prompt adherence.
On the radar
- Disney vs. Midjourney: A summary judgment deadline is approaching in November 2026 for the Disney v. Midjourney copyright case, which could set precedents for how studios license AI tools.
- Hollywood AI Deals: Reports indicate that major studios like Disney, Lionsgate, and Netflix are finalizing new deals to integrate AI into production pipelines, despite ongoing legal disputes over training data.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
