Llama-3.1-Nemotron-70B-Instruct
No estimate
No hardware requirements for this model
This model's weights are open, but no parameter count has been published for it. Every memory and speed figure starts from that number, so we would rather show nothing than a fabricated estimate.
On record
Full specification
Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.
Origin
Who built this model, where, and when it was published.
- Organisation
- NVIDIA,Meta AI
- Organisation type
- Industry,Industry
- Country
- United States of America
- Published
- 12 June 2024
- Authors
- Zhilin Wang, Yi Dong, Olivier Delalleau, Jiaqi Zeng, Gerald Shen, Daniel Egert, Jimmy Zhang, Makesh Narsimhan Sreedhar, Oleksii Kuchaiev
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Language modeling
- Approach
- Reinforcement learning
- Base model
- Llama 3.1-70B
Size
How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.
- Training data
- tokens
Training compute
The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.
- Training compute
- 7.9 × 10²⁴ FLOP
- How it was established
- Hardware
- Fine-tuning compute
- 1 × 10²⁰ FLOP
Taken from Llama 3.1 70B as the finetuning compute is multiple orders of magnitude lower
Llama 3.1 70B: 7.929e+24 FT (see Appendix F): 32+64=96 hours on a single H100 Compute: 96*60*60*989500000000000*0.3=102591360000000000000=1e20
The training run
What it physically took to train: which chips, how many, for how long, and what that drew from the wall.
- Wall-clock time
- 96 hours
Availability
Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.
- Weights
- Open — downloadable
- Model access
- Open weights (restricted use)
- Training code
- Unreleased
- Hugging Face
- nvidia
Your use of this model is governed by the NVIDIA Open Model License. Additional Information: Llama 3.1 Community License Agreement (branding restrictions + cap size of 700M MAu for commercial use). https://huggingface.co/nvidia/Llama-3.1-Nemotron-70B-Instruct-HF
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Likely above 10²³ FLOP
- Yes
- Why it is tracked
- SOTA improvement
- Record confidence
- Confident
"As of 1 Oct 2024, Llama-3.1-Nemotron-70B-Instruct performs best on Arena Hard, AlpacaEval 2 LC (verified tab) and MT Bench (GPT-4-Turbo)"
Sources
Where this record came from and when it was last checked.
- Reference
- https://www.semanticscholar.org/paper/HelpSteer2%3A-Open-source-dataset-for-training-reward-Wang-Dong/f590d8926dd12345a3bd22253461850f5ca4b3ed
- Last updated
- 28 November 2025
What the numbers mean
What this model is
Llama-3.1-Nemotron-70B-Instruct was published by NVIDIA,Meta AI, in United States of America, in June 2024. industry,Industry is the category the publisher falls under.
It works in Language, and is recorded as doing language modeling.
It is derived from Llama 3.1-70B rather than trained from scratch, which is the usual way a specialised model is produced.
Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded. It is published under the nvidia organisation on Hugging Face.
Training and provenance
Producing it required around 7.9 × 10²⁴ FLOP of arithmetic, which is a statement about the training budget rather than about inference.
The reason it appears in this catalogue at all is sOTA improvement.
Answers
Llama-3.1-Nemotron-70B-Instruct — common questions
What GPU do I need to run Llama-3.1-Nemotron-70B-Instruct?
We cannot say. Llama-3.1-Nemotron-70B-Instruct has open weights, but no parameter count has been published for it, and every memory and speed calculation starts from that number. We would rather show nothing than a fabricated estimate.
Is Llama-3.1-Nemotron-70B-Instruct open source?
Its weights are published, so Llama-3.1-Nemotron-70B-Instruct can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.
How many parameters does Llama-3.1-Nemotron-70B-Instruct have?
No parameter count has been published for Llama-3.1-Nemotron-70B-Instruct, which is why no memory or speed figure appears on this page.
Who created Llama-3.1-Nemotron-70B-Instruct?
Llama-3.1-Nemotron-70B-Instruct was published by NVIDIA,Meta AI, based in United States of America, categorised as industry,Industry.
When was Llama-3.1-Nemotron-70B-Instruct released?
Llama-3.1-Nemotron-70B-Instruct was published in June 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.
What is Llama-3.1-Nemotron-70B-Instruct used for?
Llama-3.1-Nemotron-70B-Instruct works in Language, and is recorded as handling language modeling. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
Where can I download Llama-3.1-Nemotron-70B-Instruct?
Its weights are published under the nvidia organisation on Hugging Face. We do not host model files — this site calculates what hardware is needed to run them.
How much compute was used to train Llama-3.1-Nemotron-70B-Instruct?
Around 7.9 × 10²⁴ FLOP. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.
The other direction
Looking at it from the other side?
This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.