Qwen3.6-Max-Preview
Alibaba's most powerful Qwen yet β preview flagship topping six major coding benchmarks
Qwen3.6-Max-Preview
Qwen β’ April 2026
Training Data
Up to early 2026
Qwen3.6-Max-Preview
April 2026
Parameters
Not disclosed (industry estimate ~1T sparse MoE)
Training Method
Agentic-coding focused training
Context Window
1,000,000 tokens (assumed)
Knowledge Cutoff
Not disclosed
Key Features
Preview Release β’ Agentic Programming Focus β’ QwenStudio + BaiLian API β’ Active Development
Capabilities
Coding: Outstanding
Reasoning: Outstanding
Agentic: Outstanding
What's New in This Version
+9.9 SkillsBench, +10.8 SciCode, +2.3 SuperGPQA, +5.3 QwenChineseBench vs Qwen3.6-Plus; tops SWE-benchPro, Terminal-Bench 2.0, SkillsBench, QwenClawBench, QwenWebBench, SciCode
Alibaba's most powerful Qwen yet β preview flagship topping six major coding benchmarks
What's New in This Version
+9.9 SkillsBench, +10.8 SciCode, +2.3 SuperGPQA, +5.3 QwenChineseBench vs Qwen3.6-Plus; tops SWE-benchPro, Terminal-Bench 2.0, SkillsBench, QwenClawBench, QwenWebBench, SciCode
Technical Specifications
Key Features
Capabilities
Other Qwen Models
Explore more models from Qwen
Qwen3.6-27B
Alibaba's first dense open-weight Qwen3.6 β 27B beats 397B MoE on agentic coding via hybrid Gated DeltaNet + Gated Attention
Qwen3.6-Plus
Alibaba's flagship agentic AI model with hybrid linear attention, always-on reasoning, and autonomous multi-step coding workflows
Qwen3.5-Plus
Alibaba's hosted flagship combining hybrid linear-attention MoE with native multimodal understanding for agentic workflows across 201 languages
