Meta
28 models · 8 papers in the hub
Meta’s AI research (FAIR / GenAI), best known for the open-weight Llama models that seeded the open-LLM ecosystem.
28Models
8Papers
43.7Avg index
23Open weights
2019Active since
April 8, 2026Latest
Leading model
Muse Spark
Models
- Muse Spark$0.0088.4
- Llama 4 Maverick1M$0.1563.9
- Llama 3.1 405B Instruct128K$0.8960.9
- Llama 4 Scout10M$0.0858.9
- Llama 3.1 70B Instruct131K$0.4056
- Llama 3.2 90B Instruct128K$0.3554
- Llama 3.3 70B Instruct131K$0.1050.6
- Llama 3.1 8B Instruct131K$0.0245
- Llama 3 70B Instruct8K$0.5140.8
- Llama 3.2 11B Instruct128K$0.0536.7
- Llama 3 8B Instruct8K$0.0432.4
- Llama 3.2 3B Instruct131K$0.0530.8
- Llama 2 Chat 13B$0.0028.9
- Llama 2 Chat 70B$0.0028.9
- Llama 3.2 1B Instruct131K$0.0312.1
- Llama 2 Chat 7B$0.1011.3
- Llama Guard 4 12B164K$0.18
- Llama Guard 3 8B131K$0.48
- Llama 3.3 70B128K
- Llama 3.2 11B Vision Instruct131K$0.25
- Llama 3.1 405B128K
- Llama 3 70B8K
- Llama 2 70B4K
- MusicGen
- Llama 65B$0.00
- LLaMA2K
- OPT-175B2K
- RoBERTa512
Papers
- The Llama 3 Herd of ModelsJuly 31, 2024
- Llama 2: Open Foundation and Fine-Tuned Chat ModelsJuly 18, 2023
- Segment AnythingApril 5, 2023
- LLaMA: Open and Efficient Foundation Language ModelsFebruary 27, 2023
- Toolformer: Language Models Can Teach Themselves to Use ToolsFebruary 9, 2023
- OPT: Open Pre-trained Transformer Language ModelsMay 2, 2022
- Retrieval-Augmented Generation for Knowledge-Intensive NLP TasksMay 22, 2020
- RoBERTa: A Robustly Optimized BERT Pretraining ApproachJuly 26, 2019