ASUS announced on June 22, 2026, that its XA NB3I-E12 AI server with NVIDIA HGX B300 achieved first-place rankings across key MLPerf Training v6.0 AI benchmarks. The system secured the No. 1 ranking in Llama 3.1-8B (Server and Offline modes) and GPT-OSS-120B (Interactive mode). Powered by eight NVIDIA Blackwell Ultra GPUs and dual Intel® Xeon® 6 Scalable processors, the server demonstrated superior scheduling, high-concurrency handling, maximized GPU utilization, and zero-latency processing. In the latest submission, the system achieved optimized performance in 6.584 on the Llama 2 7B training task, 84.85 on the DeepSeek 671B test, 74.75 on Llama 3.1 8B, and 2.342 on the DLRM-DCN recommendation benchmark. The server is now shipping worldwide. ASUS, a long-term member of MLCommons, participates in the MLPerf program to drive transparency and reproducibility in AI performance benchmarking.