Qwen
Creation Alibaba Cloud · ChinaThe Alibaba/Qwen family includes generation 3.8 and local and cloud models; Qwen3.8-Max with downloadable weights is not the same as that family.
| Creator | Alibaba Cloud |
|---|---|
| Access type | A combination of downloadable weights and hosted service |
| Product / Family Launch | 2023 |
| Current names and versions | Qwen3.8-Max; Qwen3.8-2.4T-A95B weights; Qwen3.5 family and Coder/VL specialized models |
| Input and output | Text, Image, Audio (in Omni models) |
| Access method | Qwen Chat, Download weights from Hugging Face, Alibaba Cloud API |
| Cost | Depending on the version: download under the same model card license, local execution with hardware cost, or API with usage cost. |
| Website | https://chat.qwen.ai |
Strengths
- Choosing between downloadable weights and hosted service
- Specialized families for text, vision, and coding tasks
Limitations
- The Max cloud version capability is not necessarily present in the released weights.
- A very large model of the family is not suitable for a regular computer; check the version and size carefully.
Best applications
Qwen is Not Just a Model
The official Qwen3.8-2.4T-A95B card confirms the release of this generation's weights. That same card describes Qwen3.8-Max as a hosted version with extra capabilities; these two should not be considered identical in terms of exact features.
Choosing for Local Execution
Instead of simply picking the biggest name, check the file size, required memory, weight format, and your execution tool's support. Read the family name, exact version, license, and quantization level together.
For a Persian project, test a few of your actual texts and evaluate extraction accuracy, writing fluency, and response speed separately.
Official sources of this review
This review was Oct 30, 2026 has been done and does not update automatically. The model, subscription plan, and product features are not identical; check your current account access and manufacturer documentation.