Meta AI

Overview

To advance AI through open research and to build open foundation models widely available to developers.

Facebook AI Research (FAIR) was established in 2013 with Turing Award laureate Yann LeCun as its founding director; it is the AI research division of Meta Platforms.

Flagship AI models

Key models released by Meta AI, most recent last. Dates are the public release month where confirmed.

1. 2023-02
   LLaMAFirst Llama research release; leaked weights spurred the open-model ecosystem
2. 2023-07
   Llama 2Open-weights model licensed for commercial use
3. 2024-07
   Llama 3.1 405BLargest openly released Llama dense model at launch
4. 2025-04
   Llama 4 (Scout / Maverick)Mixture-of-experts, natively multimodal open-weight models

Funding & financials

Quantitative figures from Epoch AI's Data on AI Companies (CC BY 4.0). Each is the latest dated figure Epoch holds; private labs disclose selectively, so undisclosed values are marked as such — never estimated.

Key facts

• Pursues an open-weights strategy: Llama models are downloadable under a community licence and are among the most-used open models worldwide.
• Long led by Chief AI Scientist Yann LeCun, co-recipient of the 2018 Turing Award for deep learning.
• In 2025 Meta reorganised its AI efforts under 'Meta Superintelligence Labs', following a large investment in Scale AI.
• Llama's licence is 'open weights' rather than OSI open source, with conditions on very large-scale commercial use.

Frequently asked

When was Meta AI founded?

Meta AI was founded in 2013, by Yann LeCun, Mark Zuckerberg. It is headquartered in Menlo Park, California, USA.

What are Meta AI's main AI models?

Meta AI's notable models include Llama 2, Llama 3.1 405B, Llama 4 (Scout / Maverick). Open-weight Llama models; FAIR fundamental research; open-model leadership.

Sources

• Meta AI — Llama 4
• Wikipedia — Meta AI
• Epoch AI — Notable AI Models
• Epoch AI — Data on AI Companies (financials, CC BY 4.0)

Other AI labs

← All AI labs