# Which AI companies release open-weight models?

> OpenAI, Meta, Google, Nvidia, Mistral, Qwen and DeepSeek each publish AI model weights, under licenses quoted here from each publisher's own page.

*The licenses range from Apache 2.0 to a publisher's own terms with conditions, and this page quotes each one from the publisher's own page.*

- Type: Explainer
- Section: Tech
- By: Tech Desk
- Published: 2026-11-20T11:19:00-05:00
- Updated: 2026-11-20T11:19:00-05:00
- Checked: 2026-11-20T11:22:00-05:00
- URL: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/

Topics: [Nvidia](https://factsdesk.com/topic/nvidia/)

## Key facts

- On Nov. 20, 2026, OpenAI's model card for gpt-oss-120b lists the Apache 2.0 license and gives 117B total parameters and 5.1B active parameters. Source: [OpenAI, on Hugging Face](https://huggingface.co/openai/gpt-oss-120b). Confirmed 2026-11-20T11:22:00-05:00. Cite: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/#on-nov-20-2026-openai-s-model-card-for-gpt-oss-120b-lists
- Meta's Llama 4 Community License Agreement, effective April 5, 2025, says a licensee whose products or services had more than 700 million monthly active users in the month before the Llama 4 release date must request a license, which Meta may grant in its sole discretion. Source: [Meta](https://dev.meta.ai/llama/llama4/license/). Confirmed 2026-11-20T11:22:00-05:00. Cite: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/#meta-s-llama-4-community-license-agreement-effective-april-5
- Meta's card for Muse-Glimmer-30B says "All artifacts are released under Apache 2.0" and lists the license as apache-2.0. Source: [Meta, on Hugging Face](https://huggingface.co/meta-models/Muse-Glimmer-30B). Confirmed 2026-11-20T11:22:00-05:00. Cite: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/#meta-s-card-for-muse-glimmer-30b-says-all-artifacts-are
- Google's Gemma 3 27B IT card lists the license as "gemma" and says a user must accept conditions to access the files. Source: [Google, on Hugging Face](https://huggingface.co/google/gemma-3-27b-it). Confirmed 2026-11-20T11:22:00-05:00. Cite: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/#google-s-gemma-3-27b-it-card-lists-the-license-as-gemma-and
- Nvidia's Nemotron 3 Nano card says use of the model is governed by the NVIDIA Nemotron Open Model License and that the model is ready for commercial use. Source: [Nvidia, on Hugging Face](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16). Confirmed 2026-11-20T11:22:00-05:00. Cite: https://factsdesk.com/tech/which-ai-companies-release-open-weight-models/#nvidia-s-nemotron-3-nano-card-says-use-of-the-model-is

As of Nov. 20, 2026, OpenAI, Meta, Google, Nvidia, Mistral AI, Qwen and DeepSeek have each published model weights, under licenses that range from Apache 2.0 to the publisher's own terms, and each license below is quoted from the publisher's own page, starting with [OpenAI's gpt-oss-120b card](https://huggingface.co/openai/gpt-oss-120b). The table covers one model per publisher, plus a second Meta model, checked on that date. It ranks nothing and does not survey every model a publisher has released.

## The license on each card

| Publisher and model | License as the page states it | Other details on the page |
| --- | --- | --- |
| OpenAI, [gpt-oss-120b](https://huggingface.co/openai/gpt-oss-120b) | "License: apache-2.0"; "Permissive Apache 2.0 license" | 117B total and 5.1B active parameters |
| Meta, [Llama 4](https://dev.meta.ai/llama/llama4/license/) | Llama 4 Community License Agreement, effective April 5, 2025 | Conditions on large licensees, naming and an acceptable use policy |
| Meta, [Muse-Glimmer-30B](https://huggingface.co/meta-models/Muse-Glimmer-30B) | "All artifacts are released under Apache 2.0" | About 29.6B parameters including a vision encoder |
| Google, [Gemma 3 27B IT](https://huggingface.co/google/gemma-3-27b-it) | "License: gemma" | A user must accept conditions to access the files; terms are linked from the card |
| Nvidia, [Nemotron 3 Nano 30B A3B](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16) | "NVIDIA Nemotron Open Model License" | "This model is ready for commercial use"; 30B total and 3.5B active parameters |
| Mistral AI, [Mistral Small 3.1 24B Instruct 2503](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503) | "Apache 2.0 License: Open license allowing usage and modification for both commercial and non-commercial purposes." | 24B parameters |
| Qwen, [Qwen3.5-27B](https://huggingface.co/Qwen/Qwen3.5-27B) | "License: apache-2.0" | 27B parameters in the language model |
| DeepSeek, [DeepSeek-V3](https://huggingface.co/deepseek-ai/DeepSeek-V3) | Code under the MIT License; models "subject to the Model License" | 671B total and 37B activated parameters per token |

The [Qwen organization page](https://huggingface.co/Qwen) says Qwen "refers to the large language model family built by Alibaba Cloud." The DeepSeek card says the "DeepSeek-V3 series (including Base and Chat) supports commercial use" and links to the Model License without reproducing its terms, so this page does not state them.

## The same licenses grouped by type

| License type on the card | Publishers and models in the table |
| --- | --- |
| Apache 2.0 | OpenAI gpt-oss-120b, Meta Muse-Glimmer-30B, Mistral Small 3.1, Qwen3.5-27B |
| The publisher's own license | Meta Llama 4, Google Gemma, Nvidia Nemotron Open Model License, DeepSeek Model License (models) alongside MIT (code) |

## The cases a reader meets next

Meta's Llama 4 agreement says that if, on the release date, the monthly active users of a licensee's products or services were "greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta," which Meta "may grant to you in its sole discretion." It also says a distributor must "prominently display" "Built with Llama," and that a model improved with Llama materials and distributed must begin its name with "Llama," according to [Meta's license page](https://dev.meta.ai/llama/llama4/license/).

Google's card says the repository "is publicly accessible, but you have to accept the conditions to access its files and content," per the [Gemma 3 27B IT card](https://huggingface.co/google/gemma-3-27b-it). The other model cards in the table show no such notice.

## What changed and when

Meta's Llama 4 license took effect April 5, 2025, per [Meta's license page](https://dev.meta.ai/llama/llama4/license/). The [Hugging Face blog](https://huggingface.co/blog/muse-glimmer) posted on Aug. 10, 2026 describes Meta's Muse Glimmer as "released under the Apache 2.0 license," and [Meta's card](https://huggingface.co/meta-models/Muse-Glimmer-30B) lists the authors as "Meta Superintelligence Lab." Meta's two models in the table therefore carry two different licenses. The gpt-oss paper listed on [OpenAI's card](https://huggingface.co/openai/gpt-oss-120b) is dated Aug. 8, 2025.

The page does not list publishers that release no weights, because no company page opened in this check states that of any company.

## Related

For what the term means and how it differs from the Open Source Initiative's definition, see [What is an open-weight AI model?](/tech/what-is-an-open-weight-ai-model/). Federal rules on chips are covered in [Commerce Department adds export license rules for advanced chips in overseas data centers](/business/ai-chip-export-restrictions/). Factsdesk reports what each page says and does not predict what any publisher or government will do.

## Sources

1. [gpt-oss-120b model card](https://huggingface.co/openai/gpt-oss-120b), OpenAI, on Hugging Face, 2026-11-20 (primary source)
2. [Llama 4 Community License Agreement](https://dev.meta.ai/llama/llama4/license/), Meta, 2025-04-05 (primary source)
3. [Muse-Glimmer-30B model card](https://huggingface.co/meta-models/Muse-Glimmer-30B), Meta, on Hugging Face, 2026-11-20 (primary source)
4. [Gemma 3 27B IT model card](https://huggingface.co/google/gemma-3-27b-it), Google, on Hugging Face, 2026-11-20 (primary source)
5. [NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 model card](https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16), Nvidia, on Hugging Face, 2026-11-20 (primary source)
6. [Mistral-Small-3.1-24B-Instruct-2503 model card](https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503), Mistral AI, on Hugging Face, 2026-11-20 (primary source)
7. [Qwen3.5-27B model card](https://huggingface.co/Qwen/Qwen3.5-27B), Qwen, on Hugging Face, 2026-11-20 (primary source)
8. [Qwen organization page](https://huggingface.co/Qwen), Qwen, on Hugging Face, 2026-11-20 (primary source)
9. [DeepSeek-V3 model card](https://huggingface.co/deepseek-ai/DeepSeek-V3), DeepSeek, on Hugging Face, 2026-11-20 (primary source)
10. [Muse Glimmer announcement](https://huggingface.co/blog/muse-glimmer), Hugging Face blog, 2026-08-10

## Asked next

- [What is an open-weight AI model?](https://factsdesk.com/tech/what-is-an-open-weight-ai-model/): An open-weight AI model is one whose trained weights are published for anyone to download and run, which differs from open source as the OSI defines it.
- [Commerce Department adds export license rules for advanced chips in overseas data centers](https://factsdesk.com/business/ai-chip-export-restrictions/): On Nov. 16, 2026, the Commerce Department released a rule requiring licenses for advanced chips in overseas data centers serving China or Macau firms.
- [What did Jensen Huang say about open-weight AI models?](https://factsdesk.com/business/what-did-jensen-huang-say-about-open-weight-ai-models/): On July 24, 2026, in his first X post, Jensen Huang wrote that open models strengthen safety and cybersecurity, per Fortune, with quotes and dates.
- [What has Anthropic said about open-weight AI models and chip exports to China?](https://factsdesk.com/business/what-has-anthropic-said-about-open-weight-ai-models-and-chip-exports-to-china/): Anthropic has backed stronger U.S. chip export controls since 2025 and says it has never asked for an open-weights ban, quoted with dates.
- [Nvidia RTX Spark compact desktops from ASUS, Dell and MSI go on sale](https://factsdesk.com/tech/nvidia-rtx-spark/): Nvidia said RTX Spark compact desktops from ASUS, Dell and MSI became available on Nov. 18, 2026. Dell lists its model from $7,464 and ASUS at $4,499.
- [Are Chinese AI models banned in the United States?](https://factsdesk.com/tech/are-chinese-ai-models-banned-in-the-united-states/): Federal orders issued in 2026 bar DeepSeek from federal agency networks; the Congressional Research Service says they do not restrict non-governmental use.
