Workspace admins can now control exactly which AI models their team can access. This highly requested enterprise feature lets you customize your workspace to match your organization's needs, budget, and creative standards—restricted models are completely hidden from member workflows.

📦 New

• Model Access Control - workspace admins can now enable or disable specific models for their team. Perfect for agencies managing client work standards, studios controlling costs, or enterprises with compliance requirements. Restricted models are completely hidden from the Editor, keeping your team focused on approved tools. To access: admins can go to Settings > Workspace Settings and toggle models.

• Minimax 2.3 - enhanced text generation with improved coherence and speed. Delivers sharper, more contextually aware outputs while processing faster than previous versions, ideal for rapid ideation and content workflows.

🧠 Improved

• VEO 3 image-to-video now supports flexible aspect ratio parameters - transform your stills into motion with precise control over output dimensions

• Wan 2.5 duration control - set custom video lengths to match your exact project needs

• Generation queue handles concurrent tasks more intelligently, prioritizing based on resource availability

• Improved processing of credit refunds for failed generations

VEO 3.1 delivers production-quality video with synchronized audio built-in, perfect for professional storytelling and cinematic workflows on FLORA.

📦 New

• VEO 3.1 - Enhanced realism with true-to-life textures, stronger prompt adherence, and native audio generation including natural conversations and synchronized sound effects

🛠 Fixes

• Video preprocessing pipeline optimized for smoother uploads

Control your video endings with Seedance Pro's last frame feature, upscale to even higher resolutions with Magnific, and enjoy more reliable generations overall.

📦 New

• Seedance 1.0 Pro last frame - precise control over video endings

• Version number now links to updates page

🧠 Improved

• Magnific maximum size increased for higher resolution outputs

• Pre-generation timeout issues resolved

🛠 Fixes

• Email change functionality fixes

• Generation stability improvements

📦 New

• Sora 2 - The latest video-gen model from OpenAI. Create high-quality user-generated content with voiceover and music directly from a single image and prompt.

Some notes on SORA 2 - the API enforces several content restrictions:

• Only content suitable for audiences under 18 (a setting to bypass this restriction will be available in the future).

• Copyrighted characters and copyrighted music will be rejected.

• Real people—including public figures—cannot be generated.

• Input images with faces of humans are currently rejected.

Make sure prompts, reference images, and transcripts respect these rules to avoid failed generations.