Lambda Labs
MLPerf Training v6.0: Lambda delivers fastest LLM training on NVIDIA GB300 NVL72 and fastest MoE training on NVIDIA HGX B200
Jun 16, 2026
Company Website
In MLCommons' MLPerf Training v6.0 round, Lambda posted the fastest Llama 3.1 8B convergence among NVIDIA GB300 NVL72 submissions (11.59 minutes, an 18.7% improvement over its prior result) and the fastest single-node HGX B200 result for OpenAI's GPT-OSS-20B.
View Original