Home / Labs / Meta AI (FAIR & Superintelligence Labs)
#49 on the AGIX Leaderboard · score 16
Meta AI (FAIR & Superintelligence Labs)
Meta's AI research and model teams, behind the Llama open-weight models.
| Type | Frontier labs |
|---|---|
| Founded | 2013 |
| Headquarters | Menlo Park, US |
| Part of | Meta |
| Website | ai.meta.com |
Meta AI (FAIR & Superintelligence Labs) research
760 papers in OpenAlex list an author affiliated with Meta (United States) in the last 12 months. The most cited:
- MapAnything: Universal Feed-Forward Metric 3D Reconstruction; map-anything.github.io20 Mar 2026 · 61 citations
- SecAlign: Defending Against Prompt Injection with Preference Optimization19 Nov 2025 · 31 citations
- Roadmap for transforming heterogeneous catalysis with artificial intelligence16 Feb 2026 · 30 citations
- Meta Audiobox Aesthetics: Unified Automatic Assessment for Speech, Music and Sound6 Dec 2025 · 24 citations
- The Open Materials 2024 (OMat24) inorganic materials dataset and models2 Jun 2026 · 23 citations
Source: OpenAlex (CC0), institution Meta (United States), read 2 Nov 2026.
Meta AI (FAIR & Superintelligence Labs) models
Latest releases
- meta-llama/Llama-Prompt-Guard-2-86M28 Apr 2025
- meta-llama/Llama-Prompt-Guard-2-22M28 Apr 2025
- meta-llama/Llama-Guard-4-12B23 Apr 2025
- meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8-Original4 Apr 2025
- meta-llama/Llama-4-Maverick-17B-128E-Instruct-Original4 Apr 2025
- meta-llama/Llama-4-Maverick-17B-128E-Original3 Apr 2025
Most downloaded
- meta-llama/Llama-3.1-8B-Instruct4,463,539 downloads (30 days)
- meta-llama/Prompt-Guard-86M2,269,781 downloads (30 days)
- meta-llama/Llama-3.2-1B-Instruct1,826,238 downloads (30 days)
- meta-llama/Llama-3.2-3B-Instruct995,167 downloads (30 days)
- meta-llama/Llama-3.1-8B550,531 downloads (30 days)
Source: Hugging Face Hub, organisation meta-llama, read 2 Nov 2026.
People at Meta AI (FAIR & Superintelligence Labs)
Similar labs
Public benefit corporation building the Claude models, with a focus on AI safety research.
91 research roles · 38 SPV filings · 975 papers#2OpenAISan Francisco, US · 2015Maker of the GPT models and ChatGPT; its charter states the mission of ensuring AGI benefits all of humanity.
73 research roles · 12 SPV filings · 375 papers · 2 models#5xAIPalo Alto, US · 2023Builds the Grok models; states its goal as understanding the true nature of the universe.
11 research roles · 24 SPV filings · 101 papers#6CohereToronto, Canada · 2019Enterprise language models (Command, Embed, Rerank); runs the Cohere Labs research group.
29 research roles · 64 papers · 25 models#7LG AI ResearchSeoul, South Korea · 2020LG's EXAONE models.
21 research roles · 118 papers · 13 models#9NVIDIA ResearchSanta Clara, USNVIDIA's research arm; publishes the Nemotron and Cosmos model families.
1359 papers · 442 modelsSomething wrong or missing? Send a correction. We fix sourced errors within a day.