Llama 3.3 Instruct 70B
MetaLlamaOpen WeightLlama 3.3 Community License Agreement
Description
Llama 3.3 is a multilingual large language model optimized for dialogue use cases across multiple languages. It is a pretrained and instruction-tuned generative model with 70 billion parameters, outperforming many open-source and closed chat models on common industry benchmarks. Llama 3.3 supports a context length of 128,000 tokens and is designed for commercial and research use in multiple languages.
Date de sortie
2024-12-06
Paramètres
70.0B
Longueur du contexte
131K
Modalités
text
Radar de capacités
26
general
18
coding
28
reasoning
36
science
80
agents
0
multimodal
Classements
| Domaine | #Rang | Score | Source |
|---|---|---|---|
| Classement codage | 525 | 17.0 | AA |
| Classement général | 405 | 34.0 | AA |
| Science | 512 | 25.0 | AA |
Scores de benchmarks (LLM Stats)
(LLM Stats (zeroeval))Chat
IFEvalGoogle Research (2023)
92.1%Aut.
General
MMLU
86.0%Aut.
BFCL v2
77.3%Aut.
Language
MMLU-Pro
68.9%Aut.
Math
MGSM
91.1%Aut.
MATH
77.0%Aut.
Reasoning
HumanEvalOpenAI (2021)
88.4%Aut.
MBPP EvalPlus
87.6%Aut.
GPQANYU + Cohere + Anthropic (2023)
50.5%Aut.
Indices d'évaluation AA
(Artificial Analysis)Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))77.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))71.3
Gpqa(NYU + Cohere + Anthropic (2023))49.8
Ifbench(Google Research (2023))47.1
Aime(MAA (Mathematical Association of America))30.0
Livecodebench(UC Berkeley + MIT + Cornell (2024))28.8
Tau2(Sierra + U Toronto + Vector Institute (2025))26.6
Lcr(Artificial Analysis)15.7
Coding Index(Artificial Analysis)11.9
Intelligence Index(Artificial Analysis)7.7
Math Index(Artificial Analysis)7.7
Aime 25(MAA (Mathematical Association of America))7.7
Terminalbench V2 14.9
Hle(Center for AI Safety + Scale AI (2025))3.6
Terminalbench Hard(Stanford × Laude Institute (2026))3.0
Scores par catégorie LLM Stats
(LLM Stats (zeroeval))Chat90
Instruction Following90
Structured Output90
Code90
Language80
Legal80
Math80
Reasoning80
Finance80
General80
Healthcare80
Tool Calling80
Physics50
Biology50
Chemistry50
Tarification
Prix d'entrée$0.71 / 1M tokens
Prix de sortie$0.72 / 1M tokens
Prix mixte (3:1)$0.712 / 1M tokens
Vitesse
Tokens/sec93.8
Délai du premier token0.62s
Temps de réponse0.62s
Classement des Prix par Fournisseur
Classement des Prix par Fournisseur
6 fournisseurs
Moins cher: DeepInfraPlus cher: Meta
FournisseurEntréeSortie
1DeepInfraMoins cher
$0
$0
2NanoGPT
$0.05
$0.23
3OpenRouter
$0.1
$0.32
4Kilo Gateway
$0.1
$0.32
5NovitaAI
$0.135
$0.4
6MetaPRINCIPAL
$0.71
$0.72
Comparer les prix entre différents fournisseurs API pour ce modèle.