1 / 3
Swipe to compare

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.

Author
AlibabaAlibaba
Release Date
2026-04-02
Knowledge Cutoff
Unknown
License
Proprietary
I/O Format
Context Length
1M / 66K
API I/O (1M)
$0.325 / $1.95
How to Use
Output Speed
53 tok/s
Arena Overall
1447
Intelligence Index
50.0
Coding Index
42.9
Math Index
LiveBench
70.8
ForecastBench
GPQA Diamond
88.2%
HLE
25.7%
MMLU-Pro
AIME 2025
MATH-500
LB Reasoning
75.8
LB Math
83.7
LB Data Analysis
69.9
LiveCodeBench
LB Coding
78.2
LB Agentic
55.0
TAU2
97.7%
TerminalBench
43.9%
SciCode
40.7%
IFBench
75.2%
AA-LCR
0.7
Hallucination (HHEM)
Factual Consistency (HHEM)
LB Language
75.0
LB Instruction Following
58.3