Is Qwen AI Open Source? Licenses, Open Weights, and Hosted Models

Last verified: August 23, 2026.

Correct answer: Do not write that “all Qwen2.5 and Qwen3 family models are Apache-2.0.” The official Qwen2.5 table assigns the Qwen Research License to 3B and the custom Qwen License to 72B. The initial downloadable Qwen3 checkpoints use Apache 2.0, but later checkpoints can use different terms. In particular, Qwen/Qwen3.8-2.4T-A95B and its official FP8 variant use the custom Qwen3.8-Max License, while Qwen/Qwen3.8-27B and its FP8 variant use Apache 2.0. A checkpoint license also does not automatically govern a hosted Max, Plus, or Flash service.^[1][15]

Qwen-AI.chat is an independent informational website. It is not operated, authorized, or endorsed by Alibaba Cloud or the Qwen Team. This page explains public Qwen model and service documentation; it does not grant a license or replace review of the applicable legal terms.

Qwen open-source status at a glance

The practical answer to “Is Qwen AI open source?” is therefore model-specific. A public download is not proof of Apache 2.0, and access through an API is not proof that the underlying weights are public.

Open source, open weights, and hosted access are different

Qwen’s own announcements have used both “open source” and “open weight.” For licensing accuracy, open weight is the safer umbrella term when an official repository publishes trained model parameters. The Open Source Initiative distinguishes open weights from a complete Open Source AI system, which also involves access to information and code needed to understand, modify, and study the system.^[20]

Model implementation code can also have a different scope from model weights. An Apache-licensed GitHub repository does not automatically assign Apache 2.0 to every checkpoint, hosted alias, training dataset, or product using that code.

Official Qwen2.5 license table

The official Qwen2.5 LLM table lists seven model sizes and does not use one license across the family.^[1] The same size-level split applies to the main base and instruction-tuned releases, but the exact repository remains the final reference.

The Qwen/Qwen2.5-3B license limits the granted rights to non-commercial research or evaluation and says commercial users must request a separate license.^[2] The 72B repository is publicly downloadable, but its custom Qwen License—not Apache 2.0—governs use.^[3]

Licenses for Qwen2.5 Coder, Math, VL, Omni, and QwQ

The Qwen2.5 name also covers specialist families. Their licenses follow the exact checkpoint, not a marketing-family label.

This corrects two common errors. First, Qwen2.5-Coder-3B is not Apache 2.0 even though the 32B Coder checkpoint is. Second, Qwen2.5-VL-72B-Instruct has public weights under the Qwen License; it should not be conflated with a separately named hosted VL-Max service.

For modality and deployment details, see our Qwen2.5-Omni guide. For the reasoning checkpoint, see QwQ-32B. The broader Qwen2.5 family guide owns general architecture and capability coverage; this page owns licensing and openness.

Which Qwen3 models are Apache 2.0?

Qwen’s April 2025 release made eight Qwen3 checkpoints available as open weights under Apache 2.0: six dense models and two mixture-of-experts models.^[15] The official Qwen3 repository states that its open-weight models use Apache 2.0.^[16]

Later Qwen3.x releases include checkpoints under more than one license. Qwen/Qwen3.5-397B-A17B and Qwen/Qwen3.6-35B-A3B identify Apache 2.0. Within Qwen3.8 itself, Qwen/Qwen3.8-2.4T-A95B and its FP8 variant identify the custom Qwen3.8-Max License, while Qwen/Qwen3.8-27B and its FP8 variant identify Apache 2.0.^[17][18] The publish-safe rule remains: use the license shown in the exact first-party repository.

See the Qwen3 model guide for architecture, thinking modes, and model selection. License verification should remain checkpoint-specific.

Why Max, Plus, and Flash should be treated separately

Hosted IDs such as qwen3.8-max, qwen3.8-27b, qwen3.7-max, qwen3.7-plus, and qwen3.7-flash are governed by provider service terms.^[19] Qwen separately publishes Qwen3.8 checkpoints under repository IDs, with a custom license for 2.4T-A95B and Apache 2.0 for 27B. Do not substitute an API ID for a repository ID, or a checkpoint license for hosted-service terms.

The same distinction applies to Qwen2.5-Max. Its official January 2025 announcement directs users to Qwen Chat and the Alibaba Cloud API but does not announce a downloadable Qwen2.5-Max checkpoint.^[22] That is evidence for hosted access, not an Apache 2.0 weight release.

A public checkpoint can also be offered through Model Studio. In that case, the two routes remain legally distinct: downloading the weights follows the repository license, while calling the managed endpoint follows the platform’s service terms. See the Qwen API guide for hosted integration and the Qwen API pricing guide for provider costs.

The same boundary applies to Qwen3.8. QwenCloud’s production qwen3.8-max is a hosted service; the retired qwen3.8-max-preview ID now routes to production rather than remaining a separate live preview. Qwen separately published 2.4T-A95B checkpoints on August 12 under the custom Qwen3.8-Max License and 27B checkpoints on August 14 under Apache 2.0. QwenCloud then launched the hosted qwen3.8-27b ID on August 19. Matching family names do not merge the service terms and repository licenses.^{[23][25][26][27][29]}

What each Qwen license means for commercial use

Apache 2.0 is permissive, but “permissive” does not mean “without conditions.” It also does not grant ownership of the Qwen trademark, guarantee non-infringement, remove regulatory duties, or make model output automatically safe to use.

How to verify any Qwen model before deployment

1. Record the exact identifier. For example, Qwen/Qwen2.5-3B-Instruct and Qwen/Qwen2.5-7B-Instruct do not share the same license.
2. Use an official Qwen repository. A community quantization or fine-tune can introduce additional files or terms.
3. Open the actual LICENSE file. Do not rely only on a search snippet, family name, blog summary, or neighboring model.
4. Check the model card and repository revision. License metadata or files can be corrected, so record the verification date.
5. Trace derivatives to the parent model. A GGUF, AWQ, adapter, or fine-tune still inherits obligations from the underlying checkpoint.
6. Separate weights from services. If the identifier is a Model Studio API alias, review provider terms rather than assuming an Apache weight release exists.
7. Review the intended use. Commercial distribution, training another model, high-scale deployment, regulated data, and trademark use can trigger different obligations.

Official sources

1. Official Qwen2.5 LLM model and license table
2. Qwen2.5-3B Qwen Research License
3. Qwen2.5-72B-Instruct Qwen License
4. Official Qwen2.5-Coder license table
5. Qwen2.5-VL-3B-Instruct license
6. Qwen2.5-VL-7B-Instruct official repository
7. Qwen2.5-VL-32B-Instruct official repository
8. Qwen2.5-VL-72B-Instruct Qwen License
9. Qwen2.5-Omni-3B license
10. Qwen2.5-Omni-7B Apache 2.0 license
11. Qwen2.5-Math-1.5B-Instruct official repository
12. Qwen2.5-Math-7B-Instruct official repository
13. Qwen2.5-Math-72B-Instruct official repository
14. Official QwQ-32B release and license statement
15. Official Qwen3 open-weight release
16. Official Qwen3 repository and license statement
17. Qwen3.5-397B-A17B official repository
18. Qwen3.6-35B-A3B official repository
19. Alibaba Cloud Model Studio model catalog
20. Open Source AI Definition 1.0
21. Apache License 2.0
22. Official Qwen2.5-Max announcement
23. QwenCloud model changelog: production Qwen3.8 Max release
24. QwenCloud text-generation model directory
25. QwenCloud retired Qwen3.8 Max preview routing
26. Qwen3.8-2.4T-A95B model card and custom license
27. Qwen3.8-27B model card
28. Qwen3.8-27B Apache 2.0 license
29. QwenCloud hosted qwen3.8-27b card

This page was last cross-checked against the official sources above on August 23, 2026.

Frequently asked questions