AI Generation
8 AI pipelines for creating richly stylized 360° panoramas — from diffusion models to instruction-based editing and style transfer
AI Generation Interface Preview
360 Hextile integrates 8 powerful AI pipelines, each optimized for 360-degree panorama generation through our unique hexagonal tiling system. From traditional diffusion models to instruction-based editing and artistic style transfer — choose the right tool for every creative goal.
Available Pipelines
SDXL Inpaint
Stable Diffusion XL with native inpainting support. Rich detail and coherent compositions at 1024x1024 native resolution. Excellent for atmospheric environments and layered visual depth.
SD 3.5 Large
Stability AI's newest architecture with improved prompt following and superior text rendering. Available in Medium, Large, and Large Turbo variants. MMDiT transformer architecture.
SD 3.5 Turbo
Distilled version of SD 3.5 optimized for speed. Generates images in just 4 steps while maintaining impressive quality. Perfect for rapid iteration and real-time workflows.
SD 1.5 Inpaint
The largest community ecosystem with 8 ControlNet types and thousands of compatible LoRAs. Low VRAM requirements make it accessible on most NVIDIA GPUs.
FLUX.1 Schnell
Ultra-fast 4-step generation from Black Forest Labs using flow matching. CFG-free distilled architecture means zero negative-prompt overhead. CLIP + T5-XXL dual text encoders with 512-token budget.
FLUX.2 Klein
Compact 4B transformer from Black Forest Labs with Qwen3 text encoder for superior prompt understanding. Two variants: Klein (distilled, 8 steps) and Klein Base (undistilled, 20 steps for best LoRA quality). GGUF Q4/Q8 and BitsAndBytes NF4 quantization for running on 8-16GB VRAM cards.
SD 2.x Inpaint
Mid-range option with 768px native resolution and 4 ControlNet types. Good depth-of-field rendering and solid general-purpose quality.
Qwen-Image
Instruction-based editing powered by Qwen2.5-VL. Describe changes in natural language — generate new panoramas or edit existing ones with two distinct modes. ControlNet Union support for structural conditioning.
TeleStyle
Diffusion-based artistic style transfer. Provide a reference style image and TeleStyle applies that aesthetic while preserving the structural content of your panorama.
Post Process
Effects-only pipeline with zero AI inference. Apply sharpening, color adjustments, blur, and other post-processing effects instantly — no GPU required.
Generation Capabilities
Image-to-Image
Transform existing panoramas while preserving structure. Adjust denoise strength to control how much the AI modifies your source image.
Inpainting
Selectively regenerate masked regions while keeping the rest intact. Perfect for fixing artifacts or replacing unwanted elements.
ControlNet
Guide generation with depth maps, edge detection, or pose estimation. Maintain spatial consistency across your entire panorama.
How Hextile AI Generation Works
Unlike traditional tiling that creates visible seams, 360 Hextile uses a hexagonal projection system that processes overlapping tiles and blends them seamlessly. Each tile is processed by your chosen AI model, then merged using GPU-accelerated alpha blending.
- Image is projected onto hexagonal tiles with 50% overlap
- Each tile is processed by the selected diffusion model
- Tiles are blended back using smooth alpha gradients
- Result is perfectly seamless at all boundaries including poles
Hextile System Diagram
Generation Examples
See what's possible with each model
SDXL - Atmospheric Forest
SD 3.5 Large - Urban Skyline
SD 3.5 Turbo - Abstract Landscape
SD 1.5 - Fantasy Environment
Qwen-Image - Instruction Edit
TeleStyle - Watercolor Style
Post Process - Sharpened Output
Format Converter - Cubemap Export
Which Pipeline Is Right for You?
Choose based on your creative goals and hardware
Low VRAM (4-8 GB)
Use SD 1.5 (6 GB) or SD 2.x (8 GB) — smaller models with rich ControlNet ecosystems.
General Purpose (12 GB)
Use SDXL — proven quality, rich LoRA ecosystem, ControlNet + IP Adapter support.
Best Quality (16+ GB)
Use SD 3.5 Large for vivid detail, or Qwen-Image for instruction-based control.
Fast Iteration (12 GB)
Use SD 3.5 Turbo — 4-step generation for rapid previews and real-time workflows.
Style Transfer (13 GB)
Use TeleStyle — apply any artistic style while preserving panorama structure.
Utilities (0-1 GB)
Use Post Process for effects-only passes, or Format Converter for projection conversions.
Model Comparison
| Model | Quality | Speed | VRAM | Best For |
|---|---|---|---|---|
|
SDXL Inpaint
|
Excellent | Medium | ~12 GB | Final production, rich environments, detailed scenes |
|
SD 3.5 Large
|
Excellent | Slower | ~16 GB | Complex prompts, text rendering, vivid detail |
|
SD 3.5 Turbo
|
Good | Very Fast | ~12 GB | Rapid iteration, previews, real-time workflows |
|
FLUX.1 Schnell
|
Good | Very Fast | ~12 GB | Ultra-fast 4-step generation, flow matching |
|
FLUX.2 Klein
|
Excellent | Medium | ~13 GB | Strong prompt adherence, compact 4B model |
|
SD 1.5 Inpaint
|
Good | Fast | ~6 GB | Low VRAM setups, ControlNet heavy workflows |
|
SD 2.x Inpaint
|
Good | Medium | ~8 GB | Mid-range option, depth-of-field rendering |
|
Qwen-Image
|
Excellent | Slower | 13-48 GB | Instruction-based editing, natural language control |
|
TeleStyle
|
Good | Medium | ~12 GB | Artistic style transfer, reference-based styling |
|
Post Process
|
N/A | Instant | 0 GB | Effects-only passes, no AI inference |
|
Format Converter
|
N/A | Fast | ~1 GB | Projection format conversion (equirect, cubemap) |
Technical Specifications
| Supported Models | SDXL, SD 3.5, FLUX.1 Schnell, FLUX.2 Klein, Qwen-Image, TeleStyle, SD 1.5, SD 2.x, Post Process, Format Converter |
| Generation Modes | Image-to-Image, Inpainting, ControlNet, Instruction Editing, Style Transfer |
| Tile System | Hexagonal projection with 50% overlap |
| Native Tile Resolution | 1024x1024 (SDXL, SD3.5, FLUX.1, FLUX.2) |
| Output Resolution | Up to 8192x4096 |
| GPU Requirements | NVIDIA GPU with 4-48 GB VRAM (pipeline dependent), CUDA 12.1+ |
| Scheduler Options | Euler, DPM++, DDIM, LMS, and more |
| Memory Optimization | xformers attention, model offloading |
| Availability | Included |
Ready to Create?
Start building expressive 360° panoramic worlds today.