Agents

Agentic inference wastes GPUs on KV cache thrashing. ThunderAgent fixes it at the scheduler level: 2.5x higher single-node throughput and ~1…

Agentic inference wastes GPUs on KV cache thrashing. ThunderAgent fixes it at the scheduler level: 2.5x higher single-node throughput and ~10x lower P50 latency at high concurrency. ThunderAgent was a

DGX agentx-post
agentstogether-ai--x

Agentic inference wastes GPUs on KV cache thrashing. ThunderAgent fixes it at the scheduler level: 2.5x higher single-node throughput and ~10x lower P50 latency at high concurrency. ThunderAgent was accepted to ICML 2026 as a Spotlight paper. Full deep-dive 👇

Source: Together AI (X) | 2026-07-29

Loading related sources…