Qwen3.8 Models by Alibaba
Last refreshed 2026-09-02. Next refresh: weekly.
Capabilities
About
Qwen3.8 is Alibaba's August 2026 Qwen generation. It includes Qwen3.8-Max, a 2.4T-total / 95B-active MoE flagship on the public DashScope/QwenCloud API, and Qwen3.8-27B, a distinct Apache 2.0 dense vision-language model with official Hugging Face weights and a public Model Studio/QwenCloud API. Compare it for Coding, Agents, Long context, and Vision.
Decision facts
- Best fit
- agentcodingvision and multimodal work
- Capability starting point
- Qwen3.8-Max-0902 with 1m context and reasoning, JSON / Tool use, structured outputs, and multimodal inputs
- Lowest tracked input
- Qwen3.8-27B · $0.5/1M · Alibaba Cloud PAI-EAS
- Closest related family
- Tongyi DeepResearch
Current Variants
Use-when guidance is based on each model's tracked capabilities, context window, release date, and replacement status.
Use when the workload needs 1m context, reasoning, and JSON / Tool use.
Use when the workload needs 262k context, reasoning, and multimodal inputs.
Use when the workload needs 1m context, 27B parameters, and reasoning.
Use when the workload needs 1m context, reasoning, and JSON / Tool use.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Qwen3.8-Max-0902 | Use when the workload needs 1m context, reasoning, and JSON / Tool use. | 2026-09 | 1m contextreasoningJSON / Tool use | Current |
| Qwen3.8-Flash-Next | Use when the workload needs 262k context, reasoning, and multimodal inputs. | 2026-08 | 262k contextreasoningmultimodal inputs | Current |
| Qwen3.8-27B | Use when the workload needs 1m context, 27B parameters, and reasoning. | 2026-08 | 1m context27B parametersreasoning | Current |
| Qwen3.8-Max | Use when the workload needs 1m context, reasoning, and JSON / Tool use. | 2026-08 | 1m contextreasoningJSON / Tool use | Current |
Release Timeline
2 release groupsSpecifications(4 models)
| Model | Released | Context | Parameters | Vision | Multimodal | Reasoning | JSON / Tool use | Structured Outputs |
|---|---|---|---|---|---|---|---|---|
| Qwen3.8-Max-0902 | 2026-09 | 1m | 2.4T | Yes | Yes | Yes | Yes | Yes |
| Qwen3.8-Flash-Next | 2026-08 | 262k | 125B total, 6B active (+51B n-gram embedding, 4B MTP) | Yes | Yes | Yes | No | No |
| Qwen3.8-27B | 2026-08 | 1m | 27B | Yes | Yes | Yes | Yes | Yes |
| Qwen3.8-Max | 2026-08 | 1m | 2.4T total, 95B active | Yes | Yes | Yes | Yes | Yes |
Available From(1 provider)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Qwen3.8-27B | Alibaba Cloud PAI-EAS | $0.5 | $3 | Serverless |
| Qwen3.8-Max | Alibaba Cloud PAI-EAS | $2 | $6 | Serverless |
| Qwen3.8-Max-0902 | Alibaba Cloud PAI-EAS | $2 | $6 | Serverless |
Related Model Families
Models(4)
Qwen3.8-Max-0902
Qwen3.8-Flash-Next
Qwen3.8-27B
Qwen3.8-Max