Llama is Meta's open-weights model family. Llama 1 (Feb 2023, research-only), Llama 2 (Jul 2023, first commercial release), Llama 3 (Apr 2024), Llama 3.1 (Jul 2024 with 405B), Llama 3.3 70B (Dec 2024). The model files are downloadable; the license is the Llama Community License, not an OSI-approved open-source license. Llama matters because it is the most-downloaded open-weights family by a wide margin and the practical foundation for most open AI work in 2023-2025. The Open Source Initiative's published position is that Llama is NOT open source under their definition, citing the license's field-of-use discrimination (the 700M-MAU clause that requires Meta's separate license for large competitors), the unilateral acceptable-use policy, and the no-competing-models clause. Compared to siblings: Mistral (smaller models historically Apache 2.0, flagships drifting to restrictive terms), Qwen (often Apache 2.0 across sizes, very open posture), DeepSeek (MIT, open posture), OLMo (truly open including data and code, smaller capability ceiling). Production-ready and dominant in deployment count. The honest framing: "open weights" is necessary but not sufficient for sovereign AI; the question is which license you actually got. Llama's license is conditionally permissive and Meta-controlled, not durably open. For the operative analysis of which clauses actually bind in your situation, see the OSI's annotated take.
The Stack · Weights · Source available
Llama (Meta)
Meta's open-weights family; dominant in usage; license carries a 700M-MAU clause and acceptable-use restrictions.
Sources
- Llama at Meta AI https://ai.meta.com/llama/
- Llama Community License Text https://www.llama.com/llama3/license/
- OSI: Meta's LLaMa license is still not Open Source https://opensource.org/blog/metas-llama-license-is-still-not-open-source
- Llama 3 Herd of Models (Meta AI, 2024) https://arxiv.org/abs/2407.21783
- ai.meta.com (audit-verified) https://ai.meta.com/blog/large-language-model-llama-meta-ai/
- about.fb.com (audit-verified) https://about.fb.com/news/2023/07/llama-2/
- ai.meta.com (audit-verified) https://ai.meta.com/blog/meta-llama-3/
- huggingface.co (audit-verified) https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct
- opensource.org (audit-verified) https://opensource.org/blog/metas-llama-license-is-still-not-open-source
Want a follow-up? Ask the chat about Llama (Meta) in context. It will compare to siblings at the same layer and ground every claim in the wiki.
Other projects at the Weights layer
9 siblings · ordered open first
- Mistral / Mixtral Open source
French lab; older open releases under Apache 2.0; flagships increasingly API-only or under research-tier licenses.
- Qwen (Alibaba) Open source
Alibaba's aggressive open-weights series (Qwen 2.5 / 3); Apache 2.0 across most sizes; full-precision weights available.
- DeepSeek V3 / R1 Open source
Cost-quality reset; V3 papers documented architectural innovations (MoE, MLA, aux-loss-free MoE); R1 open reasoning model.
- OLMo (AI2) Open source
The only major model family meeting the strictest reading of OSAID: data (Dolma), training code, and weights all published.
- Phi (Microsoft) Open source
Small open models heavy on synthetic-data training; MIT license; cost-effective inference at edge sizes.
- Kimi (Moonshot AI) Open source
Chinese open-weights series; emphasis on long-context performance.
- GLM (Zhipu AI) Open source
Tsinghua-spinoff; ChatGLM and GLM-4 families; Apache 2.0 for major releases.
- Yi (01.AI) Open source
Kai-Fu Lee's Chinese open model family (Yi-34B etc.); Apache 2.0.
- Gemma (Google) Source available
Google's open-weights siblings to Gemini; source-available, not OSI-approved.