Nvidia Unveils AVO Architecture for Long-Running AI Tasks

AI agents must retain context, plan actions and recover from errors to handle work that extends well beyond a single prompt. Nvidia’s new AVO architecture addresses those constraints with persistent memory and supervisory controls, allowing models to carry out complex, multi-step tasks over longer periods. The approach underscores a growing industry view that the systems built around a foundation model can be as important as the model itself in advancing autonomous AI capabilities.

Nvidia said AVO, paired with Anthropic’s Claude Opus 5, achieved a perfect score on the ARC-AGI-3 reasoning benchmark. The result suggests that long-term state management and execution oversight can materially improve performance without relying solely on a larger underlying model. Nvidia has not disclosed a commercial release schedule or pricing for the architecture, leaving its near-term availability and deployment model unclear.