Qwen3.6 Plus
AlibabaQwenProprietary
Description
Qwen3.6 Plus is Alibaba's next-generation flagship model featuring a 1 million token native context window, up to 65,536 output tokens, and always-on chain-of-thought reasoning. It uses a next-generation hybrid architecture optimized for efficiency and scalability. It leads on Terminal-Bench 2.0 agentic coding (61.6), surpassing Claude 4.5 Opus, and achieves strong results on document understanding (OmniDocBench 91.2) and multimodal reasoning (MMMU 86.0). Compared to Qwen 3.5, it is significantly more decisive in reasoning, using fewer tokens on straightforward tasks with better agent stability.
Date de sortie
2026-04-02
Paramètres
—
Longueur du contexte
1.0M
Modalités
image, text, video
Radar de capacités
45
general
43
coding
88
reasoning
59
scienceest.
60
agents
90
multimodal
Science utilise un proxy de raisonnement lorsque les benchmarks scientifiques dédiés ne sont pas disponibles.
Classements
| Domaine | #Rang | Score | Source |
|---|---|---|---|
| Agents & Tools | 44 | 58.0 | LS |
| Code Ranking | 31 | 78.0 | AA |
| General Ranking | 15 | 88.0 | AA |
| Multimodal Ranking | 14 | 87.0 | LS |
| Reasoning | 28 | 82.0 | LS |
| Science | 47 | 73.0 | AA |
Scores de benchmarks (LLM Stats)
Agents
WideSearch
74.3%Aut.
MCP Atlas
74.1%Aut.
TAU3-Bench
70.7%Aut.
OSWorld-Verified
62.5%Aut.
TIR-Bench
61.6%Aut.
Terminal-Bench 2.0
61.6%Aut.
Claw-Eval
58.7%Aut.
SWE-Bench Pro
56.6%Aut.
MCP-Mark
48.2%Aut.
SkillsBench
45.7%Aut.
VITA-Bench
44.3%Aut.
DeepPlanning
41.5%Aut.
Toolathlon
39.8%Aut.
NL2Repo
37.9%Aut.
Biology
GPQA
90.4%Aut.
Chemistry
SuperGPQA
71.6%Aut.
Code
SWE-Bench Verified
78.8%Aut.
SWE-bench Multilingual
73.8%Aut.
Finance
MMLU-Pro
88.5%Aut.
MMLU-ProX
84.7%Aut.
General
MMLU-Redux
94.5%Aut.
IFEval
94.3%Aut.
C-Eval
93.3%Aut.
Global PIQA
89.8%Aut.
MMMLU
89.5%Aut.
MAXIFE
88.2%Aut.
LiveCodeBench v6
87.1%Aut.
MMMU
86.0%Aut.
Include
85.1%Aut.
MMStar
83.3%Aut.
MMMU-Pro
78.8%Aut.
IFBench
74.2%Aut.
SimpleVQA
0.67 / 100Aut.
LongBench v2
62.0%Aut.
NOVA-63
57.9%Aut.
Grounding
RefCOCO-avg
0.94 / 100Aut.
ScreenSpot Pro
68.2%Aut.
Healthcare
VideoMMMU
84.0%Aut.
Language
WMT24++
84.3%Aut.
Long Context
MLVU
86.7%Aut.
AA-LCR
68.3%Aut.
MMLongBench-Doc
0.62 / 100Aut.
Math
HMMT 2025
96.7%Aut.
AIME 2026
95.3%Aut.
HMMT25
94.6%Aut.
We-Math
89.0%Aut.
DynaMath
88.0%Aut.
MathVision
88.0%Aut.
HMMT Feb 26
87.8%Aut.
IMO-AnswerBench
83.8%Aut.
PolyMATH
77.4%Aut.
Humanity's Last Exam
28.8%Aut.
Multimodal
V*
96.9%Aut.
AI2D
94.4%Aut.
OmniDocBench 1.5
91.2%Aut.
Video-MME
84.2%Aut.
CC-OCR
83.4%Aut.
CharXiv-R
81.5%Aut.
Reasoning
CountBench
0.98 / 100Aut.
ERQA
65.7%Aut.
Spatial Reasoning
RealWorldQA
85.4%Aut.
Vision
ODinW
51.8%Aut.
Indices d'évaluation AA
Intelligence Index50.0
Coding Index42.9
Tau21.0
Gpqa0.9
Ifbench0.8
Lcr0.7
Terminalbench Hard0.4
Scicode0.4
Hle0.3
Scores par catégorie LLM Stats
Video90
Biology90
Language90
Spatial Reasoning80
Structured Output80
Text-to-image80
Vision80
Chemistry80
Finance80
Frontend Development80
General80
Grounding80
Healthcare80
Instruction Following80
Legal80
Math80
Multimodal80
Physics80
Reasoning80
Code70
Economics70
Image To Text70
Long Context70
Search70
Tool Calling60
Agents60
Coding50
Tarification
Prix d'entrée$0.5 / 1M tokens
Prix de sortie$3 / 1M tokens
Prix mixte (3:1)$1.125 / 1M tokens
Vitesse
Tokens/sec52.7 tokens/s
Délai du premier token1.69s
Temps de réponse107.01s
Fournisseurs disponibles
(Unités internes LS)| Fournisseur | Prix d'entrée | Prix de sortie |
|---|---|---|
| Together | 500K | 3.0M |