HunyuanImage-3.0-Instruct

It is a native multimodal autoregressive model designed for image generation, supporting both text-to-image and image-to-image (TI2I) tasks. It features a unified architecture for multimodal understanding and generation, achieving performance comparable to leading closed-source models. The model includes two main variants: HunyuanImage-3.0 (text-to-image) and HunyuanImage-3.0-Instruct (enhanced with reasoning capabilities for intelligent prompt improvement and creative editing).

Key Features:

• Unified Multimodal Architecture: Integrates text and image modalities within an autoregressive framework, enabling contextually rich and semantically accurate image generation.
• Largest Open-Source Image Generation MoE Model:  Total parameters: 80 billion (13 billion active per token). Employs Mixture of Experts (MoE) with 64 experts.
• Supported Tasks:
  • Text-to-Image (T2I): Generates images from text prompts.
  • Image-to-Image (TI2I): Edits or modifies existing images based on instructions (e.g., "modify logo style using reference image").
  • Multi-Image Fusion: Combines multiple input images into coherent outputs.
• Intelligent Reasoning:  HunyuanImage-3.0-Instruct includes CoT (Chain-of-Thought) reasoning and prompt self-rewrite capabilities to enhance sparse or vague prompts. Enables context-aware editing (e.g., adding/removing elements, style transfers).
• Hardware Requirements:  GPU: ≥8×80GB VRAM 

Announce Date: 28.01.2026
Parameters: 80B
Experts: 64
Activated at inference: 13B
Developer: Tencent
Transformers Version: 4.50.0
vLLM Version: 0.16.0
License: tencent-hunyuan-community

Use our pre-built public endpoints for free to test inference and explore HunyuanImage-3.0-Instruct 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.