Retour à Nemotron
NVIDIA's Nemotron-tuned 70B instruct model — RLHF-optimised Llama 3.1 that tops HelpSteer2 human preference rankings.
131K tokensGratuit / Poids ouvertsTransformerLlama 3.1 Community
Benchmarks
Llama-3.1-Nemotron-70B-Instruct-HF
87.4%Nemotron-ToolOrchestra (GPT-5+Claude Opus 4.1+Qwen2.5-Math, NVIDIA)
