Qwen (Alibaba)
Qwen is Alibaba Group's family of large language and multimodal models, developed by Alibaba Cloud.
| Model | Released | Weights | Hosts | Cheapest host | Input | Output |
|---|---|---|---|---|---|---|
| Qwen3.8 Max Prime | Sep 2026 | — | 1 | Alibaba | $4.00 | $12.00 |
| Qwen3.8 Omni Flash | Sep 2026 | — | 1 | Alibaba | $0.15 | $0.47 |
| Qwen3.8 Flash | Aug 2026 | — | 3 | Alibaba | $0.15 | $0.47 |
| Qwen3.8 27B−25% | Aug 2026 | Open · Apache 2.0 | 16 | Chutes | $0.039 | $1.25 |
| Qwen3.8 2.4T A95B−20% | Aug 2026 | Open · Qwen3.8-Max License | 7 | Novita AI | $2.00 | $6.00 |
| Qwen3.8 Max | Aug 2026 | — | 2 | Novita AI | $2.00 | $6.00 |
| Qwen3.7 Flash | Jul 2026 | — | 1 | Alibaba | $0.03 | $0.13 |
| Qwen3.7 Plus | Jun 2026 | — | 1 | Alibaba | $0.40 | $1.60 |
| Qwen3.7 Max | May 2026 | — | 2 | Novita AI | $1.25 | $3.75 |
| Qwen3.6 Flash | Apr 2026 | — | 1 | Alibaba | $0.25 | $1.50 |
| Qwen3.6 27B | Apr 2026 | Open · Apache 2.0 | 9 | Chutes | $0.25 | $2.00 |
| Qwen3.6 Max Preview | Apr 2026 | — | 1 | Alibaba | $1.30 | $7.80 |
| Qwen3.6 35B A3B | Apr 2026 | Open · Apache 2.0 | 12 | AkashML | $0.10 | $0.90 |
| Qwen3.6 Plus | Apr 2026 | — | 1 | Alibaba | $0.50 | $3.00 |
| Qwen3.5 122B A10B | Feb 2026 | Open · Apache 2.0 | 5 | SiliconFlow | $0.26 | $2.08 |
| Qwen3.5 9B | Feb 2026 | Open · Apache 2.0 | 5 | SiliconFlow | $0.10 | $0.15 |
| Qwen3.5-27B | Feb 2026 | Open · Apache 2.0 | 7 | SiliconFlow | $0.25 | $2.00 |
| Qwen3.5-35B-A3B | Feb 2026 | Open · Apache 2.0 | 8 | DeepInfra | $0.14 | $1.00 |
| Qwen3.5-Flash | Feb 2026 | — | 1 | Alibaba | $0.10 | $0.40 |
| Qwen3.5 Plus 2026-04-20 | Feb 2026 | — | 1 | Alibaba | $0.40 | $2.40 |
| Qwen3.5 397B A17B | Feb 2026 | Open · Apache 2.0 | 10 | DeepInfra | $0.45 | $3.00 |
| Qwen3 Coder Next−40% | Feb 2026 | Open · Apache 2.0 | 4 | Parasail | $0.12 | $0.80 |
| Qwen3 Max Thinking | Jan 2026 | — | 1 | Alibaba | $0.78 | $3.90 |
| Qwen3 VL 8B Thinking | Jan 2026 | Open · Apache 2.0 | 1 | Alibaba | $0.18 | $2.10 |
| Qwen3 VL 32B Instruct | Oct 2025 | — | 1 | Alibaba | $0.10 | $0.42 |
| Qwen3 VL 8B Instruct | Oct 2025 | Open · Apache 2.0 | 2 | Alibaba | $0.12 | $0.46 |
| Qwen3 VL 30B A3B Thinking | Oct 2025 | Open · Apache 2.0 | 2 | SiliconFlow | $0.29 | $1.00 |
| Qwen3 VL 30B A3B Instruct | Oct 2025 | Open · Apache 2.0 | 4 | DeepInfra | $0.15 | $0.60 |
| Qwen3 Omni 30B A3B Instruct | Sep 2025 | — | 1 | Novita AI | $0.25 | $0.97 |
| Qwen3 Omni 30B A3B Thinking | Sep 2025 | — | 1 | Novita AI | $0.25 | $0.97 |
| Qwen3 Max | Sep 2025 | — | 2 | Alibaba | $1.20 | $6.00 |
| Qwen3 VL 235B A22B Instruct | Sep 2025 | Open · Apache 2.0 | 5 | DeepInfra | $0.20 | $0.88 |
| Qwen3 VL 235B A22B Thinking | Sep 2025 | Open · Apache 2.0 | 2 | Alibaba | $0.40 | $4.00 |
| Qwen3 Next 80B A3B Thinking | Sep 2025 | Open · Apache 2.0 | 2 | $0.15 | $1.20 | |
| Qwen3 Coder 30B A3B Instruct | Jul 2025 | Open · Apache 2.0 | 4 | Novita AI | $0.07 | $0.27 |
| Qwen3 30B A3B Thinking 2507 | Jul 2025 | Open · Apache 2.0 | 1 | Alibaba | $0.20 | $2.40 |
| Qwen3 Coder Flash | Jul 2025 | Open · Apache 2.0 | 1 | Alibaba | $0.30 | $1.50 |
| Qwen3 235B A22B Thinking 2507 | Jul 2025 | Open · Apache 2.0 | 3 | Alibaba | $0.23 | $2.30 |
| Qwen3 Coder Plus | Jul 2025 | Open · Apache 2.0 | 1 | Alibaba | $1.00 | $5.00 |
| Qwen3 Coder | Jul 2025 | Open · Apache 2.0 | 5 | DeepInfra | $0.30 | $1.00 |
| Qwen3 235B A22B Instruct 2507−75% | Jul 2025 | Open · Apache 2.0 | 10 | GMICloud | $0.0875 | $0.35 |
| Qwen3 32B | Apr 2025 | Open · Apache 2.0 | 2 | DeepInfra | $0.08 | $0.28 |
| Qwen3 8B | Apr 2025 | — | 1 | Alibaba | $0.18 | $0.70 |
| Qwen3 14B | Apr 2025 | Open · Apache 2.0 | 3 | NextBit | $0.10 | $0.22 |
| Qwen3 30B A3B−55% | Apr 2025 | Open · Apache 2.0 | 7 | StreamLake | $0.0481 | $0.19 |
| Qwen2.5 VL 72B Instruct | Feb 2025 | — | 1 | Parasail | $0.80 | $1.00 |
| Qwen3 Next 80B A3B | Dec 2024 | Open · Apache 2.0 | 5 | Alibaba | $0.0975 | $0.78 |
| Qwen2.5 Coder 32B Instruct | Nov 2024 | — | 1 | Cloudflare | $0.66 | $1.00 |
| Qwen2.5 7B Instruct | Oct 2024 | — | 1 | Phala | $0.10 | $0.20 |
| Qwen 2.5 72B | Sep 2024 | Open · Qwen License Agreement | 2 | DeepInfra | $0.36 | $0.40 |
| Qwen Plus 0728 | Jan 2024 | — | 1 | Alibaba | $0.40 | $1.20 |
12 of 51 models · prices per million tokens at the cheapest host, read Oct 23, 09:51 UTC
Deals now
- DiscountQwen3.7 Plusat Alibaba · until no end date
- −75%Qwen3 235B A22B Instruct 2507at GMICloud
- −55%Qwen3 30B A3Bat StreamLake
- −40%Qwen3 235B A22B Instruct 2507at StreamLake
- −40%Qwen3 Coder Nextat StreamLake
- −25%Qwen3.8 27Bat DeepInfra
- −21%Qwen3.8 27Bat Io Net
- −20%Qwen3.8 2.4T A95Bat Venice
About Qwen
- Founders
- Jack Ma (former English teacher) and 17 co-founders, including Eddie Wu (technology director at the start)3, 5
- First model
- Tongyi Qianwen, Apr 20232
- Official API operated from
- Customer-selected region: China (Beijing, Hong Kong), Singapore, Japan, Germany, US9
Alibaba Group, the parent of the Qwen family, was founded in 1999 by 18 people led by Jack Ma, a former English teacher.3, 4, 5
Its main operations are based in Hangzhou, China, and co-founder Eddie Wu has been chief executive since September 2023.3, 4, 5
The Qwen team states that it aims at artificial general intelligence and now builds generalist language and multimodal models.10
Alibaba Cloud unveiled its large language model Tongyi Qianwen on 11 April 2023 and opened a beta for enterprise customers in China.2
Its first open-source models, Qwen-7B and Qwen-7B-Chat, followed in August 2023 on ModelScope and Hugging Face, and Qwen-72B and Qwen-1.8B arrived in December 2023.11, 12
The Qwen2 series was released in June 2024, and about 100 Qwen2.5 open models followed in September 2024.11
Qwen3, announced on 29 April 2025, put several models out as open weights, including six dense models under the Apache 2.0 license.6, 7, 8
Alibaba launched the Qwen app, a personal AI assistant, in November 2025, and Qwen3.5 followed in February 2026, starting with an open-sourced model.3, 5
As of October 2026 the Qwen team presents Qwen3.8 as its latest series, described as the first release of a Qwen-Max-class model in open form.6, 7, 8
Larger Max models have otherwise been offered through Alibaba Cloud's API and Qwen Chat, and the API runs in several regions where the customer selects the data storage location.6, 7, 8, 9
Official links
- Website qwen.ai
- API docs alibabacloud.com/help/en/model-studio/what-is-model-studio
- Pricing alibabacloud.com/help/en/model-studio/model-pricing
- GitHub github.com/QwenLM/Qwen
- Hugging Face huggingface.co/Qwen
- Blog qwen.ai/research
- X x.com/Alibaba_Qwen
- LinkedIn linkedin.com/company/qwen
Sources
- Qwen - GitHub organization · github.com · read Oct 23
- Alibaba Cloud Unveils New AI Model to Support Enterprises’ Intelligence Transformation - Alibaba Group · alibabagroup.com · read Oct 23
- About Alibaba - Alibaba Group · alibabagroup.com · read Oct 23
- FAQs - Corporate Information - Alibaba Group · alibabagroup.com · read Oct 23
- Alibaba Group Holding Ltd - Form 20-F - FY2026 · sec.gov · read Oct 23
- Qwen3: Think Deeper, Act Faster · qwenlm.github.io · read Oct 23
- QwenLM/Qwen3.8 - GitHub · github.com · read Oct 23
- Qwen3-Max: Just Scale it · qwen.ai · read Oct 23
- Regions and endpoints - Alibaba Cloud Model Studio · alibabacloud.com · read Oct 23
- About Us - Qwen · qwenlm.github.io · read Oct 23
- Alibaba's Open-Source AI Journey: Innovation, Collaboration, and Future Visions - Alibaba Cloud Community · alibabacloud.com · read Oct 23
- QwenLM/Qwen - GitHub · github.com · read Oct 23
Every fact above links to its source and the date it was read; last checked Oct 23. Not affiliated with Qwen. Corrections or removal on request: contact.