M

Meta AI

Open Source

Meta AI develops the Llama family of open-source large language models — the most widely used foundation models for self-hosted AI deployment. Llama 3.3 70B and Llama 3.1 405B are the most capable open-weight models available. Unlike proprietary models, Llama can be run entirely on your own infrastructure, making it the default choice for teams with strict data privacy requirements or who want to avoid per-token API costs.

2

AI Models

128K tokens

Max Context

1

Ecosystem Tools

Current Models

Llama 3.3 70BLATEST
Context:128K tokens
Best for:High-quality open-weight reasoning with GPU self-hosting
Speed:Moderate
Pricing:Self-hosted / ~$0.60/M via API / Self-hosted / ~$0.60/M via API
Llama 3.1 405B
Context:128K tokens
Best for:Frontier-quality open-weight model for research and fine-tuning
Speed:Slow
Pricing:Self-hosted / ~$3.00/M via API / Self-hosted / ~$3.00/M via API

Meta AI Ecosystem Tools

L

LangChain

Free (OSS) / LangSmith from $39/mo

8.3

Open-source framework for building production LLM applications

ReviewVisit →

Stay in the loop

Weekly AI tool reviews, news digests, and how-to guides.