Amap Says ABot-World-0 Runs 24 Hours on One Consumer GPU
Amap, Alibaba Group's location based services platform, said in a press release that its interactive world model ABot-World-0 can sustain up to 24 hours of inference on a single consumer grade GPU. The company said this extends the model from the roughly one minute limit seen in mainstream interactive world models.
Amap said the result is based on LongForcing, a training method introduced with ABot-World-0. The method trains the model to keep image quality, physical consistency, and interaction responsiveness stable across long generation runs.
The model uses autoregressive video generation, where each new segment depends on earlier visual context. Amap said LongForcing feeds model outputs back into training while using a teacher model with longer temporal context to reduce drift across repeated generation steps.
ABot-World-0 remains open source and is available on Hugging Face and Reactor. Amap said the model is used for interactive video, game design, and simulation training.
We hope you enjoyed this article.
Consider subscribing to one of our newsletters like Daily AI Brief.
Also, consider following us on social media:
More from: AI in Game Development
Subscribe to Daily AI Brief
Daily report covering major AI developments and industry news, with both top stories and complete market updates
Market report
AI’s Time-to-Market Quagmire: Why Enterprises Struggle to Scale AI Innovation
The 2025 AI Governance Benchmark Report by ModelOp provides insights from 100 senior AI and data leaders across various industries, highlighting the challenges enterprises face in scaling AI initiatives. The report emphasizes the importance of AI governance and automation in overcoming fragmented systems and inconsistent practices, showcasing how early adoption correlates with faster deployment and stronger ROI.
Read more