Llama 3.1 70B
Meta AI's Llama 3.1 70B is a multilingual large language model optimized for dialogue use cases, leveraging an optimized transformer architecture with supervised fine-tuning and reinforcement learning with human feedback. It is genuinely best at handling multilingual text-based conversations, supporting eight languages including English, Spanish, and Hindi.
Input
Output
Context
131K
Max Output
4K
Parameters
70.6B
Technical Specifications
Model TypeChat
Context Window131,072 tokens
Max Output Tokens4,000 tokens
Parameters70.6B
Release DateJul 14, 2024
Training CutoffDec 1, 2023
Licensellama3.1
Open SourceYes
Input Modalities
Text
Output Modalities
Text
Capabilities
Resources & Links
Estimated GPU Requirements for Llama 3.1 70B
Estimates based on INT8 quantization. Actual requirements vary by framework and configuration.
Related Tools
Data sourced from official provider APIs and documentation
Last updated: Jun 24, 2026
Start building with the right model.
Automatically route workloads to the right model for every task, every time.