Llama-3 (8B) is a powerful open-source language model developed by Meta. It is part of the Llama family of models, which includes larger versions like Llama-3 (70B). Llama-3 (8B) is a pretrained and instruction-tuned generative text model optimized for dialogue use cases. It outperforms many available open-source chat models on common industry benchmarks while prioritizing helpfulness and safety.
Llama-3 (8B) is intended for commercial and research use in English. The instruction-tuned models are suitable for assistant-like chat, while the pretrained models can be adapted to various natural language generation tasks.
Llama-3 (8B) primarily supports the English language. However, since it is an open-source model, it may be possible to fine-tune or adapt it for other languages.
Llama-3 (8B) uses an optimized transformer architecture with Grouped-Query Attention (GQA) for improved inference scalability. The model has 8 billion parameters and is designed to be efficient and performant.
Llama-3 (8B) was trained on a mix of publicly available online data, with a token count of 15 trillion and a knowledge cutoff of March 2023. The training data covers a wide range of topics and domains, ensuring the model's knowledge is comprehensive and up-to-date.
Llama-3 (8B) achieves state-of-the-art results on several benchmarks:
.webp)
The model also demonstrates impressive speed, with an output speed of 119.6 tokens per second and lower latency compared to the average. It has a context window of 8,000 tokens.
Meta has developed ethical guidelines for the use of Llama-3 (8B) to ensure it is used responsibly and ethically. These guidelines cover topics such as data privacy, bias mitigation, and content moderation.
Llama-3 (8B) is released under a custom commercial license. Developers can use the model for commercial and research purposes, with specific terms and conditions outlined in the license agreement. The company's commitment to open-source and responsible AI development sets a new standard for the industry.
OpenAI-compatible — swap the base URL and it works with your existing SDK.
| Type | Price |
|---|---|
| Input | |
| Output | |
| Model | Input | Output | Context | Best for |
|---|---|---|---|---|
Llama 3 (8B) This page |
Llama 3 (8B) is priced at input $0.21 / 1M, output $0.21 / 1M.
Yes, Llama 3 (8B) can stream responses as they are generated.
Yes, Llama 3 (8B) supports both function calling and structured outputs.
Llama 3 (8B) was built by Meta.
No. The listed capabilities are function calling, streaming, and structured outputs only, with no image or video input support.
It is accessed via the v1 endpoint.
