作者： Jakub Rusinowski · 最后更新： 2026年7月10日
NVIDIA's specialized fine-tune of Llama 3.1 70B. It tops many leaderboards for helpfulness and instruction following, significantly outperforming the base Llama model in alignment.
| Nemotron 70B Instruct | Min 40 GB VRAM · Q4_K_M · 128,000 ctx · ollama run nemotron |
The cheapest GPU that runs Nemotron 70B locally (min 40 GB VRAM) is the AMD Ryzen AI Max+ 395 (96 GB).
Install Ollama then run: ollama run nemotron
Minimum VRAM: 40 GB. For best results use Q4_K_M quantization.
Nemotron 70B needs about 40 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Nemotron 70B Instruct (40 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Nemotron 70B's smallest variant needs about 40 GB, which exceeds a single RTX 4090 (24 GB). Use multiple GPUs, a higher-VRAM card, or Apple Silicon with large unified memory.
Q4_K_M is the best balance of quality and VRAM for Nemotron 70B in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
Install Ollama, then run: ollama run nemotron. This downloads Nemotron 70B and starts a local, OpenAI-compatible endpoint — no internet connection is needed after the initial download.