We’re expanding Layer’s lineup again in a big way - adding a wave of new models that give game studios even more flexibility across every part of their creative pipeline. From cinematic video generation to advanced image editing and production-ready 2D rendering, this update is all about choice.
Whether you’re building new worlds, supporting live ops weeklies, or powering performance marketing campaigns, these new additions make Layer even more adaptable - letting your team mix, match, and experiment across modalities with the same seamless workflow.
More models, more creative choice
This release brings a range of new models to Layer - from high-speed video and text-to-image generation to precision editing and refinement tools. Together, they expand what’s possible across 2D and video asset production for game studios.
Video Generation
Kling 2.5 Turbo Pro: High quality cinematics meets real-time speed, perfect for game trailers and cutscenes with precision camera controls.
Up to 1080p
First-frame/last-frame support
Reference images
Wan 2.5: Exceptional for lighting and more stylized visuals, ideal for polished in-game cinematics.
Up to 1080p
First-frame/last-frame support
Reference images & videos
Ray 2 Flash: Sharp rendering combined with fluid motion, perfect for more dynamic character scenes and hero shots.
Up to 1080p & can be upscaled to 4K
Audio support
Reference images
First-frame/last-frame support
Pixverse v5: Great for more expressive character animations or stylized worlds, a strong fit for mobile UA or game content.
360p, 540p, 720p, and 1080p
Reference images
First-frame/last-frame support
Hailuo 02 (Standard & Pro): Two options for quicker generations or visual depth, excels at more realism and complex scenes.
First-frame/last-frame support
Image Generation
Recraft v3: Great for creating clean, scalable vector-style images like icons, UI elements, and flat illustrations.
Up to 5800x5800px
Reference images
FLUX Krea: Built for creative ideation and conception, Krea is good for exploring early characters, items, or game environments.
Support for 1K; 1.2K; 1.5K; 4K resolutions
Reference images
FLUX SRPO: Focused on production-ready images that stay consistent to style and structure.
Editing & Enhancement
Sedream 4.0: Versatile model for post-gen editing and enhancement, great for targeted editing and fine tuning before production.
Seedvr2 image Upscaler: Resolution upscaling with good style preservation, for taking generations to final without losing detail.
Wan 2.5: High-fidelity video with cinematic structure

Wan 2.5 isn’t just about stills anymore - the latest version brings its signature precision in lighting, structure, and color into full video generation. This makes it a powerful tool for stylized or semi-realistic game cinematics, perfect for teams who want polished, production-ready sequences without editing needed.
Kling 2.5 Turbo Pro: Cinematic motion with real-time speed

Kling has been one of the stronger performers and popular video models on Layer for a while now. The latest generation of Kling pushes video generation even further with faster motion rendering, sharper detail, and improved temporal consistency. Kling 2.5 Turbo Pro is built for cinematic workflows - perfect for creating gameplay trailers, cutscenes, and promotional videos with dynamic camera moves and lifelike animation, all at near real-time speeds.
Ray 2: Realistic motion and cinematic depth

From Luma Dream Machine, Ray 2 focuses on smooth motion and visual fidelity. It’s ideal for creating believable character movement, cutscene animation, or cinematic previsualization. For rapid iteration, Ray 2 Flash offers a lighter, faster version tuned for quick exploration and concepting.
Natural, believable character animation out of the box
Handles secondary motion cleanly
Perfect for dialogue scenes, and character focused moments
Flash version for rapid prototyping
Seedream 4.0: Next-generation AI editing

Prompt Editing is quickly becoming one of the most-used features by game studios on Layer, and Seedream 4.0 represents a new leap forward for AI-assisted editing and enhancement. It allows localized, context-aware adjustments like inpainting and background changes to fine detail corrections - all without regenerating the whole frame. Perfect for art teams who want granular control over their final visuals without leaving the Layer workflow.
Edit scenes without full re-gens
Targeted inpainting, background swaps, and fine fixes
Granular control inside the Layer workflow
Speeds up polish and final art passes
All these models are LIVE and ready to use on Layer today - and with no more gating, anyone can jump in right now.
What’s next on Layer’s #AI features?
Advanced video refinement section:
Expand, resize, or reframe existing images or videos beyond their original borders
Big upgrade for video upscaling SeedVR2 is a more recent version that aims to do “one-step” restoration (i.e. fewer diffusion steps) through adversarial post-training, adaptive attention windows, etc