Nous Research releases open-source coding model NousCoder-14B, trained in 4 days on 48 B200 GPUs
Nous Research has released its programming model NousCoder-14B, which outperforms many proprietary systems. The model was trained in 4 days on 48 Nvidia B200 GPUs and achieved 67.87% accuracy on the LiveCodeBench v6 test, 7.08 percentage points higher than Alibaba's base model Qwen3-14B. The release coincides with the buzz around Anthropic's Claude Code tool.
Nous Research
Alibaba/Qwen
Anthropic
VentureBeat AI27.07 · 17:03



























