First AA-AgentPerf results: NVIDIA Vera Rubin NVL72 delivers up to 5.7x throughput per MW vs GB300
ArtificialAnlys · x · 2026-10-16
Artificial Analysis published the first AA-AgentPerf results for NVIDIA's Vera Rubin NVL72, the rack-scale successor to Blackwell with 72 Rubin GPUs and 36 Vera CPUs.
On agentic coding workloads, it delivers up to 5.7x the total throughput per megawatt of its predecessor GB300 NVL72 on DeepSeek V4 Pro (190 tok/s per user), 5.4x on DeepSeek V4.1 Flash (580 tok/s), 3.0x on GLM-5.2 (165 tok/s), and 2.8x on Kimi K3 (67 tok/s). It also sets the fastest Kimi K3 leaderboard result at 255 tok/s per user.
AA-AgentPerf replays real coding-agent trajectories from public repos at various concurrency levels, with inputs up to 132K tokens, short tool-call replies, and realistic tool-call pauses—reflecting how agentic traffic differs fundamentally from chat benchmarks. The full article details NVIDIA's serving techniques.
Related event: Vera Rubin NVL72 hits 5.7x GB300 throughput per MW in first tests(3 posts)→
More from coding & agent
- Burkov: Fragmented Enterprise Knowledge Breaks RAG — A Replication Agent Approach — burkov · 2026-10-16
- Marc Launvibe-codes open-source beehiiv clone MailCheap, profitable from day 1 at $19/mo — marclou · 2026-10-16
- SWE-at-Scale from SJTU synthesizes coding-agent benchmarks by inverting real bug fixes — SJTU · 2026-10-16
- JevAgentHarness: UIUC trains 4B decision models to offload LLM agent choices, cutting cost — illinois · 2026-10-16
- I built a cross-client MCP server for multi-agent rooms — here's what fought me — mcfearsome · 2026-10-16
- No moat in coding models: they'll be sold at chip cost within years, industry voices argue — rbhar90 · 2026-10-16