AGIBOT publishes LWD fleet-scale reinforcement learning for deployed generalist robots

Key points

• AGIBOT Finch published Learning while Deploying as a fleet-scale offline-to-online reinforcement learning framework for continual post-training of generalist VLA policies.
• The official page evaluates LWD on a fleet of Agibot G1 dual-arm robots across eight real-world manipulation tasks and reports consistent success-rate and cycle-time improvements over prior post-training baselines.
• It matters because this is a strong first-party research publication showing AGIBOT using deployed robot fleets as a training loop instead of treating deployment as a static endpoint.