惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

Stack Overflow Blog
Stack Overflow Blog
Vercel News
Vercel News
Cyber Security Advisories - MS-ISAC
Cyber Security Advisories - MS-ISAC
J
Java Code Geeks
M
MIT News - Artificial intelligence
Microsoft Azure Blog
Microsoft Azure Blog
B
Blog RSS Feed
MongoDB | Blog
MongoDB | Blog
G
Google Developers Blog
Engineering at Meta
Engineering at Meta
量子位
S
SegmentFault 最新的问题
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
A
About on SuperTechFans
P
Proofpoint News Feed
Last Week in AI
Last Week in AI
Recent Announcements
Recent Announcements
腾讯CDC
I
InfoQ
F
Fortinet All Blogs
Hugging Face - Blog
Hugging Face - Blog
Blog — PlanetScale
Blog — PlanetScale
H
Help Net Security
爱范儿
爱范儿

Hugging Face - Blog

ALTK‑Evolve: On‑the‑Job Learning for AI Agents Safetensors is Joining the PyTorch Foundation Holo3: Breaking the Computer Use Frontier Any Custom Frontend with Gradio's Backend A New Framework for Evaluating Voice Agents (EVA) Bringing Robotics AI to Embedded Platforms: Dataset Recording, VLA Fine‑Tuning, and On‑Device Optimizations One-Shot Any Web App with Gradio's gr.HTML CUGA on Hugging Face: Democratizing Configurable AI Agents New in llama.cpp: Model Management Building Deep Research: How we Achieved State of the Art OVHcloud on Hugging Face Inference Providers 🔥 20x Faster TRL Fine-tuning with RapidFire AI Building for an Open Future - our new partnership with Google Cloud Aligning to What? Rethinking Agent Generalization in MiniMax M2 Building a Healthcare Robot from Simulation to Deployment with NVIDIA Isaac Sentence Transformers is joining Hugging Face! Unlock the power of images with AI Sheets Supercharge your OCR Pipelines with Open Models Google Cloud C4 Brings a 70% TCO improvement on GPT OSS with Intel and Hugging Face Get your VLM running in 3 simple steps on Intel CPUs Nemotron-Personas-India: Synthesized Data for Sovereign AI Introducing RTEB: A New Standard for Retrieval Evaluation Accelerating Qwen3-8B Agent on Intel® Core™ Ultra with Depth-Pruned Draft Models VibeGame: Exploring Vibe Coding Games Nemotron-Personas-Japan: ソブリン AI のための合成データセット Swift Transformers Reaches 1.0 – and Looks to the Future Smol2Operator: Post-Training GUI Agents for Computer Use SyGra: The One-Stop Framework for Building Data for LLMs and SLMs Gaia2 and ARE: Empowering the community to study agents Scaleway on Hugging Face Inference Providers 🔥
Waypoint-1.5: Higher-Fidelity Interactive Worlds for Ever...
Andrew Lapp, Louis Castricato, Scott Fox, Shahbuland Matiana, Da · 2026-04-09 · via Hugging Face - Blog

Back to Articles

Waypoint-1.5 Weights on the Hub

Try it

waypoint 1.5

What is Waypoint-1.5?

Waypoint-1.5 is Overworld’s next real-time video world model, built to bring interactive generative worlds to the hardware people actually own.

The first release of Waypoint showed that real-time generative worlds were possible. It proved that interactive world models could be more than passive video demos, and that locally runnable systems could begin to close the gap between generating a world and actually stepping into one.

Waypoint-1.5 builds directly on that foundation. This release improves visual fidelity, expands the range of hardware that can run the model locally, and takes another step toward interactive world simulation without datacenter-scale compute.

On desktop hardware including RTX 3090 through 5090, Waypoint-1.5 can generate real-time environments at up to 720p and 60 FPS. This release also introduces a 360p tier designed to run smoothly across a much broader range of consumer hardware, including gaming laptops, and (soon) Apple Silicon Macs.

What’s new in Waypoint-1.5?

The biggest change in Waypoint-1.5 is accessibility.

With Waypoint-1, we proved the core experience. With Waypoint-1.5, we wanted to make that experience available on more machines without giving up real-time interactivity. That meant building two model tiers: a 720p model for higher-performance hardware, and a 360p model optimized for broader deployment.

We also scaled training dramatically. Waypoint-1.5 was trained on nearly 100x more data than Waypoint-1, which significantly improves the model’s ability to generate more coherent environments and more consistent motion over time.

Under the hood, Waypoint-1.5 also incorporates more efficient video modeling techniques to reduce redundant computation across frames. That matters because real-time world models are not judged only by how a single frame looks. They are judged by whether the world responds instantly, stays coherent as you move through it, and remains usable on local hardware.

Why this matters for world models

A lot of recent progress in generative video and world models has focused on visual fidelity. Those results matter, but fidelity alone is not what makes an interactive world feel real.

What people remember is responsiveness. They remember whether the environment reacts to them, whether motion stays coherent, whether the world holds together as they explore it, and whether the whole experience feels immediate instead of delayed.

That is the gap we care about most: the difference between watching a generated scene and actually being inside one.

If world models only run on large GPU clusters, they remain impressive demos. If they run locally on consumer hardware, they become something much more useful: a foundation for interactive entertainment, creative tooling, simulation, and AI-native environments people can actually explore.

Waypoint-1.5 is designed around that idea: not just better videos, but more responsive and explorable worlds that remain accessible on consumer hardware.

How to experience Waypoint-1.5

There are two ways to play Waypoint-1.5.

The first is local execution through Overworld Biome. This release is designed to run across a wide range of hardware configurations, and the updated Biome runtime makes local setup much simpler. With the new installer flow, users can go from download to running the model locally in minutes.

The second is Overworld Stream, which lets you try Waypoint-1.5 instantly in the browser with no local setup required.

Whether you want immediate access or full local control, Waypoint-1.5 is built to support both.

Additionally, we provide World Engine, our flexible, easy to use, core inference library powering our official clients, along with nearly a dozen third party clients and libraries.

The path forward

Waypoint started with a simple question: what would it take for generative worlds to become truly interactive?

Early generative systems showed that models could produce convincing images and videos. But building environments that people can explore, control, and interact with in real time is a different challenge entirely.

Waypoint-1.5 is another step in that direction, improving fidelity and expanding hardware accessibility while continuing to push real-time interactive generation onto local machines.

We think the future of world models will not be defined only by what they can render, but by whether people can actually inhabit and interact with them in real time.

Download Waypoint-1.5, run it locally with Biome, or jump in instantly on Overworld.stream.

And if you build something fun, strange, or unexpectedly immersive with it, we’d love to see it.

Stay in touch