MLCommons Releases MLPerf Training v6.1 Results, Spotlighting Recent Advances in AI Training
MLCommons has released the MLPerf Training v6.1 benchmark suite results, featuring 153 submitted systems from 24 organizations and introducing two new benchmarks: DLRMv4-HSTU for recommendations and Qwen3.5-397B for large-scale post-training workloads. The round saw an increase in cloud-based and large-scale deployments, with participants utilizing hardware from providers including NVIDIA, AMD, and Google. First-time submitter Crusoe joined established vendors in the latest benchmarking cycle. Open article →
- New AI training performance benchmarks for Google TPUs
- NVIDIA Sweeps MLPerf Training and Endpoints as Vera Rubin Debuts With Leading Performance
- NVIDIA Vera Rubin NVL72 Achieves up to 2x Blackwell Ultra Performance in MLPerf Training v6.1
- Technical Deep Dive into AMD MLPerf Training v6.1 Results
- Reproducing AMD MLPerf Training v6.1 Submission Result
- CoreWeave’s Multi-Rack NVIDIA Vera Rubin MLPerf v6.1 Results
- Lambda delivers faster AI training for language and image models
- CoreWeave Multi-Rack Rubin MLPerf Training Results
- New results: MLPerf Training v6.1. Two new benchmarks (a recommender and a first-of-its-kind LLM ...
- MLPerf® Training v6.1 results are out, and we're the first cloud provider to submit multi-rack @n...
- @MLCommons is at #SC26 in Chicago next week. MLPerf Training v6.1 results publish Monday, Nov 16....