FireRed-Image-Edit-1.1

An upgraded open-source general-purpose image editing foundation model, built upon the capabilities outlined in the FireRed-Image-Edit-1.0 Technical Report. This version significantly enhances identity consistency, multi-image conditioning, and domain-specialized editing performance, aligning closer to real-world creative production needs.  

Key Features:  

• State-of-the-Art Identity Consistency: Achieves open-source SOTA in preserving character identity across complex edits, ensuring subjects remain recognizable.  
• Multi-Element Fusion: Supports the seamless combination of 10+ elements using Agent-powered automatic cropping and stitching, reducing reliance on lengthy prompts.  
• Comprehensive Portrait Makeup: Enables dozens of professional beauty retouching styles, including skin tone brightening (yellow/olive), Halloween witch makeup, and creative looks via LoRA-based styling.  
• Text Style Reference: Maintains high-fidelity typography and stylized text quality comparable to closed-source solutions.  
• Professional Photo Restoration: Delivers high-quality restoration and enhancement of old photos with superior detail recovery.  

Engineering Optimizations:  

• Open LoRA Training Ecosystem: Full training code is released for custom style creation, with optimized samplers to maximize GPU efficiency for identical tasks, input sizes, and counts.  
• Extreme Speed Optimization: Features a complete acceleration suite (distillation, quantization, static compilation), achieving 4.5-second end-to-end generation with 30GB VRAM.  
• Intelligent Agent Workflow: Automates complex compositions (e.g., virtual try-on) without manual prompt engineering.  
• Universal Deployment: Native ComfyUI node support and GGUF lightweight format compatibility ensure seamless integration into production workflows.  

---

The model is a component of the image generation pipeline, consisting of:

• Text encoder: ~8.3B parameters,
• Transformer: ~20.4B parameters,
• VAE: ~126M parameters, 

Total: ~28.8B parameters

Announce Date: 02.03.2026
Parameters: 20B
Developer: FireRedTeam
Diffusers Version: 0.36.0.dev0
vLLM Version: 0.12.0rc1
License: Apache 2.0

Use our pre-built public endpoints for free to test inference and explore FireRed-Image-Edit-1.1 capabilities. You can obtain an API access token on the token management page after registration and verification.

Rent your own physically dedicated instance with hourly or long-term monthly billing.

We recommend deploying private instances in the following scenarios:

• maximize endpoint performance,
• enable full context for long sequences,
• ensure top-tier security for data processing in an isolated, dedicated environment,
• use custom weights, such as fine-tuned models or LoRA adapters.

Contact our dedicated neural networks support team at nn@immers.cloud or send your request to the sales department at sale@immers.cloud.