LAGEN research shows a 1-frame delay can cost real-time agents 95% performance
wzenus · x · 2026-10-22
Researchers introduced LAGEN (Latency-Aware Agents), studying how deployment latency affects vision-language agents acting in real time.
- Key finding: a 1-frame delay, invisible to humans, can cost an agent up to 95% of its performance on real-time tasks
- The world doesn't pause while an agent thinks; training with deployment latency helps VLM agents and VLAs act in real time
- Models like Astra already run real-time loops, but latency awareness is needed to keep them alive in real-world timing
More from Research
- Break-Glass paper lays out emergency biosecurity licensing regime for an AI biological incident — davidmanheim · 2026-10-22
- Lafa scales flow language models to 8B, matching AR quality at 2.5x throughput — AdtRaghunathan · 2026-10-22
- New paper: SMC sampling with integrator snippets yields principled HMC tuning strategies — michaelchchoi · 2026-10-22
- HKU-led team launches Embodied Playground, 410-task benchmark for model physical intelligence — YiMaTweets · 2026-10-22
- Chatlet: Generative UI That Emerges From Group Conversations — round · 2026-10-22
- NJU's Socratic Skill Learning hits 0.801 utility, beating distillation baselines — NJU-LINK · 2026-10-22