Llama-3.3-70B-Instruct

Meta · meta/llama-3-3-70b-instruct

ProviderMeta
Familyllama
Typellm-chat
Statusactive
Released2024-12-06
Updated2024-12-06

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
Aider Polyglot Benchmarkunassessed45.2
AlpacaEvalunassessed40.5
Arena Elo — Codingunassessed1250.0
Arena Elo — Hard Promptsunassessed1320.1
Arena Elo — Mathunassessed1240.0
Arena Elo — Overall (Text)unassessed1260.0
Arena Elo — Style Controlunassessed1318.0
Artificial Analysis Intelligence Indexunassessed76.0
Artificial Analysis Output Speedunassessed73.0
BIG-Bench Hardunassessed58.1
BBQ (Bias Benchmark for QA)unassessed80.1
GPQA Diamondunassessed50.2
GSM8Kunassessed95.0
HELM Safetyunassessed84.2
HumanEvalunassessed80.5
IFEvalunassessed78.5
LiveCodeBenchunassessed28.8
MATH-500unassessed73.2
MGSM (Multilingual Grade School Math)unassessed91.1
MMLU: Astronomyunassessed71.5
MMLU: Biology (subcategory)unassessed80.2
MMLU: Business Ethicsunassessed72.8
MMLU: Chemistry (subcategory)unassessed72.5
MMLU: Clinical Knowledgeunassessed78.5
MMLU: Computer Science (subcategory)unassessed80.5
MMLU: Jurisprudenceunassessed69.5
MMLU: Physics (unresolved key)unassessed76.8
MMLU-Prounassessed66.5
MMLU: Professional Accountingunassessed63.8
MMLU: Professional Lawunassessed68.2
MT-Benchunassessed8.6
MultiPL-Eunassessed72.9
MultiPL-E: C++unassessed72.8
MultiPL-E: C#unassessed72.5
MultiPL-E: Gounassessed68.5
MultiPL-E: Javaunassessed77.5
MultiPL-E: JavaScriptunassessed75.8
MultiPL-E: Juliaunassessed52.2
MultiPL-E: Kotlin (unconfirmed)unassessed62.5
MultiPL-E: Luaunassessed45.8
MultiPL-E: Perlunassessed40.5
MultiPL-E: PHPunassessed70.2
MultiPL-E: Pythonunassessed80.2
MultiPL-E: Runassessed48.5
MultiPL-E: Rubyunassessed55.8
MultiPL-E: Rustunassessed60.5
MultiPL-E: Scalaunassessed50.2
MultiPL-E: Swiftunassessed55.5
MultiPL-E: TypeScriptunassessed74.8
MuSRunassessed24.5
SWE-bench Verifiedunassessed25.8
ToxiGenunassessed88.2
WildBenchunassessed68.2

Data

This card as JSON · Edit on GitHub