Qwen 3.6 Plus
QwenFlagship
ThinkingTool UseVisionStructured Output
About this model
Qwen 3.6 Plus builds on a hybrid architecture that combines efficient linear attention with sparse mixture-of-experts routing, enabling strong scalability and high-performance inference. Compared to the 3.5 series, it delivers...
Performance Tier
Flagship
Qwen 3.6 Plus is a flagship model from Qwen : the most capable in their lineup.
Best-in-class model from this provider. Highest performance across benchmarks, ideal for demanding tasks.
Pricing
This model is included in Elosia plansAffordable
Low cost. Suitable for sustained use and high-volume interactions.
| Type | per 1M tokens |
|---|---|
| Input (prompt) | $0.325 |
| Output (completion) | $1.95 |
| Cache write | $0.406 |
Capabilities
Context Length1.0M
Max Output Tokens66K
TokenizerQwen3
Inputtext, image, video
Outputtext
Release DateApril 2, 2026
Benchmarks
General Intelligence
MMLU
Not reported
GPQA Diamond
Not reported
Mathematics
MATH-500
Not reported
Programming
HumanEval
Not reported
SWE-bench Verified
78.8%
SWE-bench Multilingual
73.8%
Reasoning
IFEval
Not reported
Agentic
SWE-bench Pro
56.6%
Terminal-Bench 2.0
61.6%
Recommended Use Cases
CodingAnalysisResearchGeneral ChatTranslation
Strengths
- Alibaba frontier flagship — state-of-the-art agentic coding (SWE-bench Verified 78.8%)
- Hybrid architecture: linear attention + sparse MoE routing for efficient 1M-token context
- Leads Terminal-Bench 2.0 (61.6%) — ahead of Claude 4.5 Opus (59.3%)
- Strong repository-level reasoning and frontend workflow fluency
- Hybrid thinking mode — toggle chain-of-thought reasoning per request
Limitations
- Alibaba's official evaluation focused on agentic and multimodal benchmarks — MMLU, GPQA, HumanEval and MATH-500 are not part of the reported suite
- Architecture details (exact parameter counts, MoE topology) not publicly disclosed
- Hosted-only — the open-weight variant Qwen3.6-35B-A3B is a different model
Resources
This model may use your data for training
