Qwen3.8 Max
Active Parameters
2.4T
Context Length
1M
Modality
Multimodal
Architecture
Mixture of Experts (MoE)
License
Proprietary
Release Date
2 Aug 2026
Knowledge Cutoff
-
Evaluation Benchmarks
Rank
#17
| Benchmark | Score | Rank |
|---|---|---|
Agentic Coding max | 0.65 | 🥇 1 |
Agentic Coding Standard | 0.65 | 🥇 1 |
Web Development max | 1669 | 🥈 2 |
General max | 0.78 | 7 |
General Standard | 0.78 | 7 |
LiveBench Average max | 0.78 | 7 |
LiveBench Average Standard | 0.78 | 7 |
Reasoning max | 0.88 | 12 |
Reasoning Standard | 0.88 | 12 |
Data Analysis max | 0.78 | 13 |
Data Analysis Standard | 0.78 | 13 |
Agent Arena Standard | 0.06 | 14 |
Agent Arena max | 0.06 | 16 |
General Text max | 1481 | ⭐ 16 |
General Text Standard | 1481 | ⭐ 16 |
Mathematics max | 0.91 | 17 |
Mathematics Standard | 0.91 | 17 |
Coding max | 0.73 | 34 |
Coding Standard | 0.73 | 34 |
Rankings
Overall Rank
#17
Coding Rank
#77
About Qwen3.8 Max
Qwen3.8-Max is Alibaba Cloud's flagship hosted foundation model based on the 2.4T parameter (95B active) Mixture-of-Experts architecture. Featuring a hybrid Gated DeltaNet linear attention and standard Gated Attention design, it provides native vision input capabilities, default 1M token context window, non-thinking and thinking mode toggles, and built-in agent tooling for complex software engineering, research reproduction, and long-horizon autonomous tasks.
Technical Specifications
Detailed architecture specifications are limited for proprietary models.
Resources
About Qwen 3.8
Alibaba's Qwen 3.8 generation represents the frontier hybrid Mixture-of-Experts architecture designed for coding, professional work, research, and long-horizon agentic tasks. It features a 2.4-trillion parameter architecture (95B active per token) combining Gated DeltaNet linear attention with standard Gated Attention, available both as open weights and as a hosted flagship service.