Llama 3.1 70B

Meta · meta/llama-3-1-70b

ProviderMeta
Familyllama
Typellm-chat
Statusactive
Released2024-07-14

Reported benchmark scores

These scores carry one collection date for the whole card (2026-04) and a source list rather than a source per score, so they cannot be attributed individually. They are shown as reported and are not verified evidence. Benchmark pages carry the reviewed, dated evidence where it exists.
BenchmarkCatalogue standingScore
Arena Elo — Codingunassessed1230.0
Arena Elo — Mathunassessed1220.0
Arena Elo — Overall (Text)unassessed1240.0
Artificial Analysis Intelligence Indexunassessed75.0
Artificial Analysis Output Speedunassessed72.0
BIG-Bench Hardunassessed55.8
BBQ (Bias Benchmark for QA)unassessed78.2
FinBenchunassessed55.2
GPQA Diamondunassessed42.5
HELM Safetyunassessed82.5
HumanEvalunassessed76.2
IFEvalunassessed76.0
LegalBenchunassessed58.5
LiveCodeBenchunassessed23.2
MATH-500unassessed64.5
MedQAunassessed65.2
MMLU: Astronomyunassessed70.2
MMLU: Biology (subcategory)unassessed79.5
MMLU: Business Ethicsunassessed71.5
MMLU: Chemistry (subcategory)unassessed71.2
MMLU: Clinical Knowledgeunassessed77.2
MMLU: Computer Science (subcategory)unassessed79.8
MMLU: Jurisprudenceunassessed68.8
MMLU: Physics (unresolved key)unassessed75.5
MMLU-Prounassessed60.8
MMLU: Professional Accountingunassessed62.5
MMLU: Professional Lawunassessed67.5
MultiPL-Eunassessed70.8
MultiPL-E: C++unassessed70.5
MultiPL-E: C#unassessed70.2
MultiPL-E: Gounassessed66.8
MultiPL-E: Javaunassessed75.8
MultiPL-E: JavaScriptunassessed73.5
MultiPL-E: Juliaunassessed50.5
MultiPL-E: Kotlin (unconfirmed)unassessed60.5
MultiPL-E: Luaunassessed43.8
MultiPL-E: Perlunassessed38.5
MultiPL-E: PHPunassessed68.5
MultiPL-E: Pythonunassessed78.5
MultiPL-E: Runassessed46.5
MultiPL-E: Rubyunassessed53.5
MultiPL-E: Rustunassessed58.2
MultiPL-E: Scalaunassessed48.2
MultiPL-E: Swiftunassessed53.2
MultiPL-E: TypeScriptunassessed72.5
MuSRunassessed22.7
ToxiGenunassessed86.5

Data

This card as JSON · Edit on GitHub