LLMs are General Asynchronous Agents
cs.LG, cs.CL
Submitted: 2026-09-28
Updated: 2026-09-28
Comments: Preprint
Code: https://github.com/dvmazur/async_llm
License: http://creativecommons.org/licenses/by/4.0/
The gist: Modern LLMs are increasingly capable as autonomous agents, but they follow sequential interaction cycles: read, think, reply or call tools, repeat.
Terminology
Abstract
Modern LLMs are increasingly capable as autonomous agents, but they follow sequential interaction cycles: read, think, reply or call tools, repeat. Many real-world use cases are not sequential: voice assistants, embodied agents, and monitoring systems receive new inputs while they think or perform another task. Modern LLMs address this with specialized architectures for voice interaction and video streams, VLAs for robot control, asynchronous tool calling for API usage, and others. In this work, we generalize from different asynchronous tasks to general asynchronous agents that can adapt to different types of concurrency. To achieve this, we develop an asynchronous LLM framework that lets users (or the agents themselves) define inference coroutines with overlapping memory states. We showcase that Qwen 3.x models are capable of asynchronous operation for streaming video understanding, videogames, and monitoring, without task-specific training.
Sources
- Laguna M.1/XS.2 Technical Report
- Gemini Robotics: Bringing AI into the Physical World
- SARATHI: Efficient LLM Inference by Piggybacking Decodes with Chunked Prefills
- Do As I Can, Not As I Say: Grounding Language in Robotic Affordances
- GQA: Training Generalized Multi-Query Transformer Models from Multi-Head Checkpoints
- Longformer: The Long-Document Transformer
- Better speech synthesis through scaling
- GR00T N1: An Open Foundation Model for Generalist Humanoid Robots
- LLM-Based Generalizable Hierarchical Task Planning and Execution for Heterogeneous Robot Teams with Event-Driven Replanning
- Interruption Handling for Conversational Robots
- MapQA: A Dataset for Question Answering on Choropleth Maps
- Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models
- Moshi: a speech-text foundation model for real-time dialogue
- Extreme Compression of Large Language Models via Additive Quantization
- DualVLA: Building a Generalizable Embodied Agent via Partial Decoupling of Reasoning and Action
- GPTQ: Accurate Post-Training Quantization for Generative Pre-trained Transformers
- Gemma 4 Technical Report
- Asynchronous LLM Function Calling
- Asynchronous Tool Usage for Real-Time Agents
- Robotouille: An Asynchronous Planning Benchmark for LLM Agents
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: The Predictive Cost of ELBO Binding
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks