Qwen 3.6 Plus is Alibaba's latest flagship language model released in April 2026, delivering significant advances in agentic coding, multimodal code generation, and 1M-token context handling. As a native multimodal model, it can generate front-end web pages from screenshots, design drafts, and text prompts, and autonomously breaks down complex programming tasks through iterative write-test-debug cycles. It matches Claude Opus 4.5 on programming and agent benchmarks including SWE-bench and Terminal-Bench 2.0, making it a strong contender for enterprise agentic AI deployment.
VisionReasoning|Proprietary Model
Knowledge Cutoff
Unknown
Input → Output Format
Context Memory
1MIN66KOUT
AI Performance Evaluation
Arena Overall Score
1447
±7As of 2026-05-01
Overall Rank
No.46
7,538 Votes
Arena by Ability
Hard Prompts
1468±9No.44
Expert Knowledge
1477±23No.38
Instruction Following
1434±12No.45
Conversation Memory
1439±17No.61
Creative
1397±17No.68
Coding
1506±13No.32
Math
1478±25No.13
Arena by Occupation
Creative Writing
1420±14No.51
Social Sciences
1445±16No.70
Media
1400±15No.61
Business
1454±16No.33
Healthcare
1472±26No.40
Legal
1448±25No.57
Software
1484±11No.45
Mathematics
1454±26No.38
Source:Arena Intelligence
Overall
AA Intelligence Index
50%↑11%
LiveBench
71%↑10%
Reasoning & Math
GPQA Diamond
88%↑6%
HLE
26%↑8%
LB Reasoning
76%↑7%
LB Math
84%↑10%
LB Data
70%↑17%
Coding
AA Coding Index
43%↑6%
LB Coding
78%↑5%
LB Agentic
55%↑10%
TAU2
98%↑17%
TerminalBench
44%↑10%
SciCode
41%↓1%
Language & Instructions
IFBench
75%↑12%
AA-LCR
70%↑8%
LB Language
75%↑3%
LB IF
58%↑7%
Output Speed
Standard Mode
53tok/s↓24
First Output 106.94s
