What are the hardware requirements for Meta Llama 3.1 405B Instruct FP8?

Meta Llama 3.1 405B Instruct FP8 requires approximately 197.6 GB of VRAM and supports a context window of 128K tokens. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed Meta Llama 3.1 405B Instruct FP8 and how large is it?

Meta Llama 3.1 405B Instruct FP8 is developed by meta-llama, a model with 405b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

How does Meta Llama 3.1 405B Instruct FP8 perform on standard benchmarks?

Meta Llama 3.1 405B Instruct FP8 has the following published scores: LMSYS ELO 1333. Compare against reference models on this page or on the LLM Explorer leaderboards.

Meta Llama 3.1 405B Instruct FP8 by meta-llama — VRAM 197.6GB, 128K context

Name: Meta Llama 3.1 405B Instruct FP8
Author: meta-llama

Meta Llama 3.1 405B Instruct FP8 is an open-source language model by meta-llama. Features: 405b LLM, VRAM: 197.6GB, Context: 128K, License: meta, Instruction-Based, LLM Explorer Score: 0.33, ELO: 1333.

Arxiv:2204.05149 Autotrain compatible Base model:meta-llama/meta-lla... Base model:quantized:meta-llam... Conversational De En Endpoints compatible Es Facebook Fbgemm fp8 Fr Hi Instruct It Llama Llama-3 Meta Pt Pytorch Region:us Safetensors Sharded Tensorflow Th

Model Card on HF 🤗: https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct-FP8

Meta Llama 3.1 405B Instruct FP8 Benchmarks

LMSys ELO 1333 vs 1272 (so35) 4.8%

^nn.n% — How the model compares to the reference models: Anthropic Sonnet 3.5 ("so35"), GPT-4o ("gpt4o") or GPT-4 ("gpt4").

What is the LLM Explorer Rank (Score)

Meta Llama 3.1 405B Instruct FP8 Parameters and Internals

LLM Name	Meta Llama 3.1 405B Instruct FP8
Repository 🤗	https://huggingface.co/meta-llama/Meta-Llama-3.1-405B-Instruct-FP8
Base Model(s)	meta-llama/Meta-Llama-3.1-405B-Instruct meta-llama/Meta-Llama-3.1-405B-Instruct
Model Size	405b
Required VRAM	197.6 GB
Updated	2024-09-25
Maintainer	meta-llama
Model Type	llama
Instruction-Based	Yes
Model Files	4.9 GB: 1-of-109 4.0 GB: 2-of-109 4.7 GB: 3-of-109 4.7 GB: 4-of-109 4.5 GB: 5-of-109 4.4 GB: 6-of-109 4.7 GB: 7-of-109 4.7 GB: 8-of-109 4.7 GB: 9-of-109 4.5 GB: 10-of-109 4.4 GB: 11-of-109 4.7 GB: 12-of-109 4.7 GB: 13-of-109 4.7 GB: 14-of-109 4.5 GB: 15-of-109 4.4 GB: 16-of-109 4.7 GB: 17-of-109 4.7 GB: 18-of-109 4.7 GB: 19-of-109 4.5 GB: 20-of-109 4.4 GB: 21-of-109 4.7 GB: 22-of-109 4.7 GB: 23-of-109 4.7 GB: 24-of-109 4.5 GB: 25-of-109 4.4 GB: 26-of-109 4.7 GB: 27-of-109 4.7 GB: 28-of-109 4.7 GB: 29-of-109 4.5 GB: 30-of-109 4.4 GB: 31-of-109 4.7 GB: 32-of-109 4.7 GB: 33-of-109 4.7 GB: 34-of-109 4.5 GB: 35-of-109 4.4 GB: 36-of-109 4.7 GB: 37-of-109 4.7 GB: 38-of-109 4.7 GB: 39-of-109 4.5 GB: 40-of-109 4.4 GB: 41-of-109 4.7 GB: 42-of-109 4.7 GB: 43-of-109
Supported Languages	en de fr it pt hi es th
Model Architecture	LlamaForCausalLM
License	meta
Context Length	131072
Model Max Length	131072
Transformers Version	4.43.0.dev0
Vocabulary Size	128256
Torch Data Type	bfloat16

Rank the Meta Llama 3.1 405B Instruct FP8 capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to Meta Llama 3.1 405B Instruct FP8

Best Alternatives	Context / RAM	Downloads	Likes
Llama 3.1 405B Instruct FP8	128K / 209.2 GB	4760	16
Llama 3.1 405B Instruct	128K / 183.1 GB	43229	595
Meta Llama 3.1 405B Instruct	128K / 186 GB	55654	473
Llama 3.1 405B Instruct FP8	128K / 193.4 GB	15734	193
Shisa V2 Llama3.1 405B	128K / 191.2 GB	22	19
BigLlama 3.1 681B Instruct	128K / 189.6 GB	17	12
INTELLECT 1 Instruct	8K / 20.5 GB	328	125
...ama 3.1 405B Instruct Bnb 4bit	128K / 214.3 GB	141	9