Llama 3.1 Instruct 70B

Last updated: Recently

Model Type

Text

Max Context

Max Output

Parameter Scale

Llama 3.1 Instruct 70B by Meta: Text model; TTFT 0.736s, 28.8 tok/s.

Input Modality

Output Modality

Inference Speed

28.769 tokens/s

Latest Release Date

7/23/2024

SDK Ecosystem

artificial-analysismanufactureraa-bootstrap

Model Overview Fields

name

Llama 3.1 Instruct 70B

Release date

7/23/2024

Performance

28.769 tokens/s

model_creator.name

Meta

AA Evaluation Scores

AA Intelligence Index

12.5

AA Coding Index

10.9

AA Math Index

MMLU Pro

0.676

GPQA

0.409

HLE

0.046

LiveCodeBench

0.232

SciCode

0.267

Math-500

0.649

AIME

0.173

AIME 2025

0.04

IFBench

0.344

LCR

0.063

TerminalBench Hard

0.03

TAU2

0.152

Output Speed (tokens/s)

28.769 tok/s

TTFT (s)

0.736s

First Answer Token (s)

0.736s