Waypoint-1.5: A Real-Time Video World Model for Consumer Hardware

2026-09-29

AI Summary

This study addresses the latency and throughput bottlenecks that interactive world models encounter when responding to dense control signals on consumer-grade hardware. To this end, it proposes an end-to-end real-time interactive video generation framework tailored for low-compute devices. Methodologically, a diffusion model is pretrained on one hundred thousand hours of gameplay data, integrated with a large-scale alignment data pipeline, specialized training strategies, and an efficient runtime system to achieve real-time playable video generation under full keyboard-and-mouse input with multi-resolution adaptation. In terms of contributions, this work pioneers an evaluation metric system that distinguishes among rendering, latent, and control frame rates, achieves real-time low-latency interaction across a broad range of consumer hardware, and establishes corresponding safety and ethical evaluation standards.

Abstract

We present Waypoint 1.5, a real-time diffusion world model for interactive video generation on consumer-grade hardware. Unlike general video diffusion models, interactive world models (iWMs) must respond to dense user controls under strict latency and throughput constraints. Waypoint 1.5 is pre-trained on 100,000 hours of diverse, control-aligned video game data across hundreds of games, and generates playable video conditioned on full keyboard and mouse input. The model includes two resolution variants that run across a wide spectrum of consumer hardware. To characterize this unique setting, we distinguish rendered FPS, latent FPS, and control rate. We describe the data pipeline, architecture, training methodology, and runtime system behind Waypoint 1.5. We evaluate interactivity through latency and throughput. Finally, we discuss the safety and ethics considerations unique to iWMs.

Problem

• interactive world models
• real-time video generation
• consumer hardware
• latency constraints
• user controls

Innovation

• Interactive World Model
• Real-time Diffusion
• Video Generation
• Consumer Hardware
• Control-aligned Data