CoreWeave Sets New AI Training Records in MLPerf® Training v6.0, Training DeepSeek-V3 in Approximately Two Minutes
CoreWeave set a new record in the MLPerf Training v6.0 benchmark by training the 671B-parameter DeepSeek-V3 model in 2.02 minutes on 8,192 NVIDIA GB300 GPUs. The results demonstrated near-linear scaling efficiency across cluster sizes, validating the performance of its Mission Control orchestration and SUNK scheduling infrastructure.
Key figures
- Gpt oss 20b time min
- 26.98
- Llama 3 1 8b time min
- 16.54
- Max gpus cluster size
- 8192
- Llama 3 1 405b time min
- 9.77
- Deepseek v3 2048 gpu time min
- 5.54
- Deepseek v3 4096 gpu time min
- 3.09
- Deepseek v3 8192 gpu time min
- 2.02
AI analysis
The rest of the AI analysis, red flags and sentiment are part of Signal8 Pro.
AI-generated analysis of a public disclosure. Not investment advice; verify against the original document.