Skip to content
TheToolsVerse LogoThetoolsverse
Submit Tool
Hunyuan GameCraft – Tencent's Open-Source Interactive Game Video Generation Model logo

Hunyuan GameCraft

Visit Hunyuan GameCraft
Pricing: Free
Verified: Yes
Editor rating: 4.4/5
Updated: July 2026

Tencent's open-source AI model that generates interactive, action-controllable game video sequences from a single image and keyboard inputs.

Editor's take: “Tencent's impressive 3D world generation model” — Sohail Akhtar

Free tier (verified July 2026): Free weights + inference code — catch: Research/non-commercial; check repo licence for commercial use

Top Alternatives
View all AI Simulation tools Browse all free AI tools
Editor-selected listing
Verified by our team
Independent & reader-supported

Editor's Verdict

Official Review
Hunyuan GameCraft is a technically significant open-source contribution to interactive game video generation from Tencent, offering keyboard-controllable scene navigation with genuine scene coherence across long sequences — a practical advancement over prior models in this category. It is a research and prototyping tool rather than a production asset pipeline, and hardware requirements for the quantized model still require a current high-end consumer GPU as a minimum.
4.4 / 5.0
Editor Rating

Reviewed by Sohail Akhtar

Lead Editor & Founder

Pros

What we like

  • Trained on over one million gameplay recordings across more than 100 AAA game titles, giving the model broad coverage of game visual styles from realistic to stylized environments within a single open-source release
  • The hybrid history-conditioned architecture meaningfully addresses scene coherence during navigation — one of the most visible quality issues in prior interactive video generation models — by preserving original scene information after camera movement
  • The quantized 13B version runnable on an RTX 4090 makes the model accessible to individual developers and researchers without requiring data center-grade hardware, unlike the full-precision version at 80GB VRAM

Cons

Limitations

  • Hunyuan GameCraft generates video sequences that simulate gameplay — it does not produce real-time interactive output, meaning it cannot replace game engine rendering for production titles and is best suited for prototyping and concept visualization
  • The full-precision model's 80GB VRAM requirement restricts deployment to high-end workstation or server hardware; even the quantized version requires an RTX 4090 or equivalent, limiting accessibility for developers without current-generation high-end GPUs

Pricing

✓ Free tier re-verified July 2026: Free weights + inference code — Research/non-commercial; check repo licence for commercial use

PlanDetails
FreeFully free and open-source. Model weights and inference code available on GitHub and Hugging Face. Full model requires approximately 80GB VRAM; quantized 13B version runs on consumer GPUs such as the RTX 4090.

Hunyuan GameCraft is free and open-source. Model weights and inference code are available at GitHub (github.com/Tencent-Hunyuan/Hunyuan-GameCraft-1.0) and Hugging Face (tencent/Hunyuan-GameCraft-1.0). No licensing fee or subscription is required for research and non-commercial use; users should review the repository license for commercial application terms.

What is Hunyuan GameCraft?

Quick Summary

Hunyuan GameCraft is an open-source AI model released by Tencent's Hunyuan team in August 2025 that generates interactive, playable-quality game video sequences from a single reference image and keyboard or mouse input signals, enabling developers and researchers to simulate navigable game environments without traditional rendering pipelines. Built on the HunyuanVideo foundation model and trained on over one million gameplay recordings across more than 100 AAA games, it is designed for game developers, technical researchers, and AI creators who want to prototype interactive game environments using generative video synthesis. Model weights and inference code are freely available on GitHub and Hugging Face.

Hunyuan GameCraft is an open-source interactive game video generation framework developed by the Tencent Hunyuan research team and released in August 2025. The model generates dynamic, playable video sequences from a single scene reference image and action inputs — keyboard signals (W, A, S, D, arrow keys) and mouse movements — that control character navigation and camera direction within the generated environment. Its core architectural innovation is a hybrid history-conditioned training strategy that preserves scene information across extended sequences as the user navigates, preventing the visual drift and inconsistency that affected earlier interactive video generation models. The framework also generates dynamic environmental content including moving clouds, rain, snow, and water flow, rather than static scene backdrops. See top alternatives. Hunyuan GameCraft is intended for game developers who want to prototype environment concepts and character interactions without committing to full 3D modeling and rendering workflows, AI researchers studying interactive generative video models, and technical creators who want to experiment with keyboard-controlled scene exploration. The model was trained on a dataset of over one million gameplay recordings spanning more than 100 AAA titles, which gives it broad coverage of game visual styles from realistic environments to stylized and pixelated aesthetics. Outputs from Hunyuan GameCraft are compatible with standard game engine workflows — generated video sequences can serve as visual references or concept validation for environments built in Unity or Unreal Engine View alternatives.
Read the full overview
The full-precision model requires approximately 80GB of GPU VRAM for optimal inference, which places it outside the range of most consumer setups. A quantized 13-billion parameter version, available through Hugging Face, can run on consumer-grade GPUs including the RTX 4090, making the model accessible to individual developers and researchers with high-end consumer hardware. See top alternatives. PCM distillation reduces inference steps and computational overhead while maintaining temporal consistency across long video sequences. The model does not generate real-time playable output — it generates video sequences that simulate gameplay, which can be used for prototyping and visualization but not as a drop-in replacement for real-time game engine rendering in production titles View alternatives.

Associated Tags

AI game video generation, open-source game AI, interactive video generation, Tencent Hunyuan AI, game environment simulation, generative game world, keyboard-controlled AI video

Key Features

Keyboard and mouse action-controlled video generation
Hybrid history-conditioned scene preservation
Trained on 1M+ gameplay recordings across 100+ AAA games
Dynamic environmental content generation
Quantized 13B model for consumer GPU deployment
Open-source weights on GitHub and Hugging Face

Target Audience

Who should use Hunyuan GameCraft?

Game developers and level designers who want to prototype environment concepts and navigation feel without full 3D productionAI and computer vision researchers studying interactive generative video synthesis and world model architecturesTechnical creators and developers with access to high-end GPUs who want to experiment with open-source game world generationIndie developers exploring AI-assisted concept visualization for game environments before committing to engine-based developmentStudents and researchers benchmarking open-source interactive video generation models for academic or exploratory projects
Real Use Cases

How professionals leverage Hunyuan GameCraft – Tencent's Open-Source Interactive Game Video Generation Model

Discover practical workflows and real-world scenarios where Hunyuan GameCraft delivers key solutions.

01

Prototyping an interactive game environment concept by generating video sequences from a single scene image and keyboard navigation inputs before investing in 3D modeling

02

Testing how a proposed character movement system feels in a specific visual environment by running keyboard action inputs through the generative model

03

Using the model as a visual reference generation tool for Unity or Unreal Engine level designers to explore scene layouts before building them in the engine

04

Researching the current state of interactive generative video models as part of an AI or computer vision research project

05

Generating stylized game world explorations from a reference concept image to share with a development team for environment direction feedback

06

Experimenting with the quantized 13B model on an RTX 4090 to evaluate the model's practical capability for an indie game development workflow

Top Alternatives

Dedicated alternatives page →
Freemium
Hunyuan World 1.0

Hunyuan World 1.0

Open-source AI model by Tencent that generates explorable, interactive 3D worlds from text or image inputs using panoramic scene reconstruction.

#3D Model#AI Simulation
View Details
Free
Matrix-Game 2.0

Matrix-Game 2.0

Skywork AI's 1.8B open-source interactive world model generating real-time 25 FPS gameplay from keyboard and mouse inputs, with long-sequence consistency and free weights on GitHub and Hugging Face.

#AI Simulation#Amazing+3
View Details
Free
Oasis AI Game

Oasis AI Game

Open-source AI world model by Decart and Etched that generates real-time Minecraft-style interactive gameplay at 20 FPS using next-frame prediction, with no traditional game engine required.

#AI Simulation#Amazing+2
View Details
Free
RTFM

RTFM

World Labs' real-time generative world model that creates persistent, navigable 3D scenes from a single image on one H100 GPU.

#AI Simulation#Games+2
View Details

Frequently Asked Questions

What is Hunyuan GameCraft?
Hunyuan GameCraft is Tencent's open-source AI model that generates interactive game video sequences from a single reference image and keyboard or mouse input signals, enabling navigation and exploration of AI-generated game environments.
Is Hunyuan GameCraft free and open-source?
Yes. Model weights and inference code are freely available on GitHub at Tencent-Hunyuan/Hunyuan-GameCraft-1.0 and on Hugging Face at tencent/Hunyuan-GameCraft-1.0. Users should review the repository license for commercial application terms.
What GPU is required to run Hunyuan GameCraft?
The full-precision model requires approximately 80GB of GPU VRAM. A quantized 13B parameter version is available that runs on consumer GPUs such as the RTX 4090.
Can Hunyuan GameCraft integrate with Unity or Unreal Engine?
Hunyuan GameCraft generates video sequences rather than real-time rendered output. These videos can be used as visual reference and prototyping assets for Unity or Unreal Engine development, but do not function as a direct engine integration for real-time gameplay.
What data was Hunyuan GameCraft trained on?
The model was trained on over one million gameplay recordings across more than 100 AAA game titles, providing broad coverage of game visual styles from realistic environments to stylized and pixelated aesthetics.
Who should use Hunyuan GameCraft?
Hunyuan GameCraft is best suited for game developers prototyping environment concepts, AI researchers studying interactive generative video, and technical creators with high-end GPU hardware who want to experiment with open-source game world generation.