Qwen
by Alibaba
At a glance
Family of open-weight large language models developed by Alibaba, spanning text, vision, and coding tasks, including the Qwen3.5 and Qwen3.8-Max lines.
llm open-weights multimodal
- Licence
- Apache-2.0 (Qwen2.5 open models except the 3B and 72B variants); later Qwen lines not read
- Category
- Open Models
- Business model
- Open (open weights/source)
- Availability
- Global
- Launched
- 2023-08
Overview
Qwen is Alibaba's open-weight large language model family, covering text, vision, coding, and reasoning variants across many sizes. Recent public releases include the Qwen3.5 line and Qwen3.8-Max. Most releases use the Apache-2.0 license, making Qwen one of the most downloaded open model families worldwide. Models are available on Hugging Face and through Alibaba Cloud's API.
Compare this tool Suggest a correction
Key features
- Broad family: text, vision-language, coder, and reasoning models
- Recent Qwen3.5 and Qwen3.8-Max releases
- Apache-2.0 licensing for most releases
- Available via Hugging Face and Alibaba Cloud API
Use cases
- Self-hosted multilingual assistants
- Code generation with Qwen coder variants
- Fine-tuning open models for vertical applications
Pricing
Open-weight Qwen models are free to download; hosted access through Alibaba Cloud is usage-priced.
Record details
Provenance, dates and links are kept here so the main sections stay about the product itself.
Links, dates and provenance
- Record updated
- 2026-10-13
- Model family
- Qwen
- Founded
- 1999
- Headquarters
- CN
- Canonical URL
- https://globalaiproductindex.com/products/qwen/
Selection profile
Every field says where its value comes from. "Unknown" means we have not confirmed it — it does not mean the product lacks the feature.
Selection profile11 fields2 unknown
- Tasks it fits
- Running models yourself
Derived from the record: category "open-models".
Suggest a correction - How you use it
- Model weights
Derived from record tag matching /^open-weights$/.
Suggest a correction - Runs on
- Not recorded
No platform support is stated in the record.
Suggest a correction - Where it runs
- Local machine
Open-weight models are run by the user, not served by the vendor.
Suggest a correction - Works offline
- Partly
Local execution implies the model can run without a network once downloaded, but the vendor does not state this in the record.
Suggest a correction - Official service region
- Global
Mirrors the availability field already on the record ("global").
Suggest a correction - Public API
- Yes
The record mentions an API surface.
Suggest a correction - Licence
- Apache-2.0 (Qwen2.5 open models except 3B and 72B; later Qwen lines not read)
The Qwen2.5 release notes state: "All our open-source models, except for the 3B and 72B variants, are licensed under Apache 2.0." The Qwen2.5-7B-Instruct model card carries license:apache-2.0. The value is scoped to Qwen2.5; later Qwen lines were not read.
Suggest a correction - Commercial use
- Permitted
Scope: Apache-2.0 permits commercial use for the Qwen2.5 models it covers; the Qwen2.5 3B and 72B variants use different terms, and later Qwen lines were not read.
Suggest a correction - Interface languages
- Not recorded
No vendor page read in this pass states the product's own interface languages; the key exists so the field can be answered later rather than left absent.
Suggest a correction - Billing
- Open source Free Free allowance: Yes Subscription required: No
Prices are the vendor’s advertised entry price and exclude taxes; always confirm on the vendor’s own page.
Suggest a correction
Where each value comes from
- Tasks it fits
- Our inference
- How you use it
- Our inference
- Runs on
- Our inference
- Where it runs
- Our inference
- Works offline
- Our inference
- Official service region
- Our inference
- Public API
- Our inference
- Licence
- Verified 2026-10-13 qwenlm.github.io ↗
- Commercial use
- Our inference
- Interface languages
- Unknown
Key facts and their evidence
Each fact lists the conditions it holds under, the page it was read from and the day it was checked.
-
Licence: Apache-2.0 for the Qwen2.5 open models, except the 3B and 72B variants Verified 2026-10-13 qwenlm.github.io ↗
Applies to: Applies to the Qwen2.5 open-source releases; the Qwen2.5-3B and Qwen2.5-72B variants are excluded, and later Qwen lines are not covered.
Verbatim from the Qwen2.5 release notes: "All our open-source models, except for the 3B and 72B variants, are licensed under Apache 2.0." The Qwen2.5-7B-Instruct model card carries license:apache-2.0.
Suggest a correction -
Licence: The Qwen2.5 3B and 72B variants are not covered by the Apache-2.0 statement Verified 2026-10-13 qwenlm.github.io ↗
Applies to: Scope statement for the Qwen2.5 line.
The release notes carve out the 3B and 72B variants and point readers to the licence file in each Hugging Face repository instead.
Suggest a correction -
Licence: Later Qwen lines are outside this record's licence reading Our inference
Applies to: Scope statement: the record names Qwen3.5 and Qwen3.8-Max elsewhere without reading their licences.
Only the Qwen2.5 licence text was read; the newer lines named in the record are not asserted to share it.
Suggest a correction
Frequently asked questions
Is Qwen open source?
It depends on the line and the size. In the Qwen2.5 release read for this record, all open-source models except the 3B and 72B variants are Apache-2.0, and they can be self-hosted and fine-tuned; the 3B and 72B variants use different terms. This record's licence reading covers Qwen2.5, not the later Qwen3.x lines, and hosted access through Alibaba Cloud is a service rather than a download.
Who develops Qwen?
Qwen is developed by Alibaba's AI teams and distributed via Hugging Face and Alibaba Cloud.
What is Qwen3.8-Max?
Qwen3.8-Max is a late-2026 Qwen flagship in Alibaba's Qwen3 line, following earlier Qwen3.5 releases for text, coding, and reasoning.
What does the Qwen record not record?
No public documentation or pricing page is recorded for Qwen, because no same-site page could be confirmed as either. What is recorded is the licence: the Qwen2.5 release notes and the Qwen2.5-7B-Instruct model card give Apache-2.0 for the Qwen2.5 open models except the 3B and 72B variants. Terms for later Qwen lines are not recorded.
Similar products
- DeepSeek — Open-weight LLM family and chat assistant from Hangzhou-based DeepSeek, known for strong reasoning models.
- Mistral — Family of open-weight large language models from Mistral AI, including the Mistral and Mixtral series.
- Baichuan — Open-weight bilingual LLM family from Beijing-based Baichuan Intelligence, now focused on medical AI with its Baichuan-M reasoning models.
- GLM — Family of open-weight large language models from Zhipu AI, spanning general, reasoning, and coding-focused variants.
- Llama — Open-weight large language model family from Meta, spanning Llama 3 and Llama 4 generations across multiple sizes.
- Yi — Open-weight bilingual (English and Chinese) large language model family from 01.AI.
Related comparisons
Related guides
Sources
- Project repository ↗
- Qwen (Wikipedia) ↗
- Qwen2.5 release notes ↗
- Qwen2.5-7B-Instruct model card (Hugging Face) ↗
This record was last reviewed on 2026-10-13.
Machine-readable record: /api/products/qwen.json · Spot an error? Suggest a correction