In the news
RadixArk Miles runs agentic RL training on NVIDIA Vera CPUs: faster sandboxes, faster training
LMSYS · Published · 3 min read
In 30 seconds
- What happened
- RadixArk Miles completed agentic RL training with all sandboxes on NVIDIA Vera CPUs, showing 12% faster episodes than AMD EPYC Turin.
- Why it matters
- Matters for engineers training large language models with reinforcement learning who need to optimize sandbox infrastructure and training throughput.
- Watch out
- Results are task-specific to Terminal-Bench and SWE-bench; gains depend on inference speed and full CPU load conditions that may not apply everywhere.
- agent
- agentic
The patterns behind this
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.