Tencent Releases Open-Source World Model Hy-Nomas
Tencent Releases Open-Source World Model Hy-Nomas
Tencent has open-sourced Hy-Nomas, a world model with text and camera control, Apache-2.0 code and weights, and LoRA fine-tuning support.
Tencent has released Hy-Nomas, a world model announced a few days before Genie 4. It joins the existing lineup of MIRA, Solaris and Odyssey-3, adding text and camera control for video generation.
- Video is generated in fragments, with context between fragments preserved in a KV cache.
- Camera movement can be specified in advance or changed between fragments.
- Code and weights are released under Apache 2.0.
- The model can be fine-tuned with LoRA on custom videos; ready-made adapters are available for cinematic visuals and robotics.
For generation on a single RTX 4090, there is a configuration with INT8 quantization, limited history and model offloading to RAM. The speed of this mode is not specified. In Reactor with full history on 8×B200, the team reported about 18.5 FPS generation.
Why it matters
AnalysisOpen-source world models with camera control and LoRA fine-tuning let developers experiment with interactive video generation, while single-GPU configurations lower the barrier to entry.
Discuss in community
Share your questions and insights with developers
Related in Ecosystem
Practical guides, core terminology, and agent skills
Open Weights vs Truly Open Source AI
A legal and engineering analysis of the fundamental difference between accessible numerical weights of models (Llama, DeepSeek) and fully open projects with source data, code, and architecture (OSI Standard).
Open-source method enables 6K image generation for Qwen Image, Ideogram, and Flux
A new open-source method supports creating images at very high resolutions, up to 6K, across several popular generation models.
Jev and System One AI: A Complete Practical Guide to the Decision Model from TypeSafe
A deep engineering breakdown of the flagship Jev model: System One architecture versus LLMs, Choice/Score/Noul primitives, request parallelism, confidence mathematics, video use cases, code, and an overview of local open-weight alternatives (Laya, GLiNER2.5, CLM-8B).
Meta Llama (The Premier Open Standard for Artificial Intelligence)
The flagship series of open language models from Meta (Llama 3, 3.3). Available for free download, modification, and local deployment on personal servers without restrictions or censorship.