Nemotron 70B

Open Source
By NVIDIAReleased 2024-10-14128K context70B parameters
Open SourceAPI Available

Benchmark Scores

78.4/100
MMLU83.5
HumanEval80
Math68.5
Reasoning80.5
Coding79.5

About Nemotron 70B

Nemotron 70B is NVIDIA's fine-tuned variant of Llama 3.1 70B, optimized using NVIDIA's RLHF pipeline. It achieves notably higher scores on the Arena Hard benchmark than the base model.

Strengths

  • +Strong instruction following
  • +NVIDIA optimized
  • +Based on Llama 3.1
  • +Good at chat

Weaknesses

  • -Derivative of Llama
  • -Limited unique capabilities
  • -Less community adoption

Pricing

Per 1M tokens

Input$0.18
Output$0.18

Quick Facts

Context Window
128K
Parameters
70B
Release Date
2024-10-14
Category
Open Source

Compare Models

See how Nemotron 70B stacks up against other models

More from NVIDIA

Try Nemotron 70B on Vincony

Compare Nemotron 70B with 400+ other AI models side-by-side. Run AI debates, generate images and videos, and find the best model for your task.

Visit Vincony.com