# Hunyuan Image 3.0 AI Generator

> Tencent’s native multimodal base generator for complex scenes and high prompt alignment.

Canonical URL: https://allimageai.com/models/hunyuan-image-3/

Website: All Image AI

Language: English

Hunyuan Image 3.0 is Tencent’s large open text-to-image base model, with 80 billion total parameters and 13 billion active parameters. This page keeps its direct diffusion controls and optional prompt expansion separate from the reasoning-oriented Instruct version.

## Model details

- Developer: Tencent
- Status on All Image AI: Active
- Last verified: 2026-08-25
- Maximum output: 2K
- Modes: Text to Image
- Reference images: not supported
- Resolutions: 1K, 2K
- Aspect ratios: 1:1, 16:9, 9:16, 4:3, 3:4, 3:2, 2:3

## Best for

- Complex scene composition
- Detailed descriptive prompts
- Open-model research workflows
- Large final renders

## How to use

1. Describe spatial relationships clearly.
2. Use a short negative prompt.
3. Test the composition at 1K.
4. Render only the winner at 2K.

## Settings explained

### Guidance

7.5 is the documented base-model default for prompt adherence.

### Prompt Expansion

Disabled by default so the expensive run follows the text you supplied.

## Prompt examples

### Crowded scene

> Telephoto photograph through a festival crowd, one laughing subject in sharp focus, layered banners and lanterns, natural background compression

### Technical cutaway

> Architectural cutaway of a compact tidal research station, labelled functional zones, coherent structure, restrained scientific illustration

## Limitations

- Substantially more expensive than distilled draft models
- Text-to-image only here
- The 2K tier should be reserved for final candidates

## Pricing and credits

Credits for one image at default quality and mode. Resolution, quality, reference inputs, edit mode and output count can change the total. Check the Generate button before submitting.

| Resolution | Credits / image |
| --- | --- |
| 1K | 8 |
| 2K | 30 |

Failed generations cost 0 credits. Catalog availability does not guarantee runtime availability; the workspace reports whether the selected model can run.

[Credit plans](https://allimageai.com/pricing.md)

## Test results

Testing in progress. No verified benchmark scores have been published for this model.

## Commercial use and image handling

Commercial use is governed by Tencent’s Hunyuan model licence and the selected service terms.

Uploads are processed by the configured inference service. You can delete a generation from your history.

## Frequently asked questions

### How is this different from Hunyuan Image 3 Instruct?

This is the base text-to-image model with classic diffusion controls; Instruct adds a reasoning-oriented prompt workflow.

### Why does 2K cost many more credits?

The upstream route bills by output megapixel, so a square 2K image contains four times the pixels of 1K.

## Release history and sources

September 28, 2025 — Weights and model code released

This date applies to the event shown. Later API access, weights and availability on All Image AI may have different dates.

- [github.com](https://github.com/Tencent-Hunyuan/HunyuanImage-3.0)

[Event evidence on the timeline](https://allimageai.com/timeline/#event-hunyuan-image-3)

## Other versions

- [Hunyuan Image 3 Instruct](https://allimageai.com/models/hunyuan-image-3-instruct.md): Prompt expansion · 1K
- [Hunyuan Image 2.1](https://allimageai.com/models/hunyuan-image-2-1.md): Steps + guidance · 1K