ThunderAgent: 2x Faster Agentic Inference for Synthetic Data Generation

togethercompute · x · 2026-07-30

Together AI introduced ThunderAgent, a high-throughput system designed for large-scale agentic inference. By introducing a novel program abstraction for scheduling agentic LLM requests, it effectively eliminates KV cache thrashing.

Key Performance Results:

Engineering Integration:

The research was accepted as a Spotlight paper at ICML 2026.

Related event: Together AI Unveils ThunderAgent for 2x Agent Inference Speedup(6 posts)→

Original post →

More from Infra

Infra channel →