LLaMA 65B
The 65B tier of the first LLaMA generation, initially offered by application for noncommercial research.
FAMILY / META
Llama
Meta’s open-weight foundation-model line, evolving from general language models to native multimodal MoE systems.
- First model
- 2023
- Models
- 10
- Latest dated version
- Llama 4 Scout
01 / RELEASE SEQUENCE
Family release sequence
- 2023-02-24LLaMA 65BLanguage
- 2023-07-18Llama 2 70BLanguage
- 2023-08-24Code Llama 34BCoding
- 2024-04-18Llama 3Language
- 2024-07-23Llama 3.1 405BLanguage · Reasoning
- 2024-09-25Llama 3.2 3BLanguage
- 2024-09-25Llama 3.2 VisionLanguage · Multimodal
- 2024-12-06Llama 3.3 70BLanguage · Reasoning
- 2025-04-05Llama 4 MaverickLanguage · Reasoning
- 2025-04-05Llama 4 ScoutLanguage · Reasoning
02 / GENERATIONS
What changed
Llama 2 70B
The 70B tier of Llama 2, available for research and commercial use under its community license.
Code Llama 34B
A coding-specialized model continued from Llama 2; 34B was the largest launch tier, with 70B added later.
Llama 3
Represents the separate 8B and 70B checkpoints released together in April 2024, without assigning one parameter count to the whole release.
Llama 3.1 405B
Llama’s first 405B open flagship, extending the family to 128K context.
Llama 3.2 3B
The lightweight text-only Llama 3.2 tier for mobile and edge devices, separate from the 11B/90B vision models.
Llama 3.2 Vision
The 11B and 90B Llama 3.2 vision tiers for image understanding, separate from the lightweight text-only branch.
Llama 3.3 70B
Brought near-Llama-3.1-405B instruction quality to a much smaller 70B model.
Llama 4 Maverick
The higher-capability Llama 4 open model with 400B total parameters and 128 experts.
Llama 4 Scout
A natively multimodal MoE with 17B active parameters, 16 experts, text-and-image input, and up to 10M context.