Llama 3.3 70B Instruct
Metaยทmeta/llama-3.3-70b-instruct
Industry standard open weights 70B model matching Llama 3.1 405B capabilities across core coding, reasoning, and multilingual benchmarks.
Open WeightsOpenTool CallingStructured OutputsNatural Language ProcessingArtificial Intelligence
Context
128,000
Output Limit
8,192
Linked Papers
64
Citations
2.9K
Release Date
Dec 6, 2024
Model Specifications
- Lab / Vendor
- Meta
- Model Family
- Meta
- Modality
- text
- Architecture
- Dense Grouped-Query Attention (GQA) Transformer
- Parameters
- 70B Dense
- Context Window
- 128,000
- Max Output
- 8,192
- Input / Output Price
- $0.18 / $0.59 per 1M
- License
- Llama 3.3 Community
Research & Evaluation Papers
Academic evaluations, empirical benchmarks, and publications for Llama 3.3 70B Instruct.
Sort