Llama-3.1-70B-Instruct

meta-llama

Hugging Face ↗

Parameters

70B

Downloads

732.9K

on HF

Likes

921

Avg score

62.5

10 benchmarks

Weights

Open

Released

Jul 16, 2024

About

Llama-3.1-70B-Instruct is a 70B-parameter open-weights model in the Llama family from meta-llama. Benchmarked on 10 evals, averaging 62.5.

llama3.1text-generation

Benchmark scores · 10

avg 62.5 — higher is better

GSM8Kcurated · 2026-05

95.1

IFEvalcurated · 2026-05

87.5

MMLUcurated · 2026-05

HumanEvalcurated · 2026-05

80.5

MATHcurated · 2026-05

BIG-Bench HardOpen LLM LB · 2026-05

55.9

GPQA Diamondcurated · 2026-05

MMLU-ProOpen LLM LB · 2026-05

47.9

MATH Lvl 5Open LLM LB · 2026-05

38.1

MuSROpen LLM LB · 2026-05

17.7

The research behind this model

all papers mentioning it →

The Carbon Footprint of Machine Learning Training Will Plateau, Then Shrink

Details

Publishermeta-llama

FamilyLlama

Parameters70B

Licensellama3.1

WeightsOpen

ReleasedJul 16, 2024

History

full ledger →

Capability over time

Append-only ledger — every observed change to this model's numbers.