Qwen3.6-Plus-2026-04-02

Model IntroductionText

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.

Introduction

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.

Key Features

  • 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.

Benchmarks

Benchmark (Area)Score*
AA-LCR (Long Context)68.3 (Eval)
AI2D (Multimodal)94.4 (Eval)
AIME 2026 (Math)95.3 (Eval)
CC-OCR (Multimodal)83.4 (Eval)
C-Eval (Reasoning)93.3 (Eval)
CharXiv-R (Multimodal)81.5 (Eval)
Claw-Eval (Agents)58.7 (Eval)
CountBench (Reasoning)1.0 (Eval)
DeepPlanning (Reasoning)41.5 (Eval)
DynaMath (Math)88.0 (Eval)
ERQA (Reasoning)65.7 (Eval)
Finance Agent v2 (Reasoning)40.8 (Eval)
FrontierSWE (Agents)22.0 (Eval)
Global PIQA (Physics)89.8 (Eval)
GPQA (Physics)90.4 (Eval)
HMMT 2025 (Math)96.7 (Eval)
HMMT25 (Math)94.6 (Eval)
HMMT Feb 26 (Math)87.8 (Eval)
Humanity's Last Exam (Math)28.8 (Eval)
IFBench (Instruction Following)74.2 (Eval)
IFEval (Chat)94.3 (Eval)
IMO-AnswerBench (Math)83.8 (Eval)
Include (General)85.1 (Eval)
LiveBench (Math)70.9 (Eval)
LiveCodeBench v6 (Reasoning)87.1 (Eval)
LongBench v2 (Long Context)62.0 (Eval)
MathVision (Math)88.0 (Eval)
MAXIFE (General)88.2 (Eval)
MCP Atlas (Reasoning)74.1 (Eval)
MCP-Mark (Agents)48.2 (Eval)
MLVU (Long Context)86.7 (Eval)
MMLongBench-Doc (Long Context)0.6 (Eval)
MMLU-Pro (Language)88.5 (Eval)
MMLU-ProX (Language)84.7 (Eval)
MMLU-Redux (Language)94.5 (Eval)
MMMLU (Language)89.5 (Eval)
MMMU (Multimodal)86.0 (Eval)
MMMU-Pro (Multimodal)78.8 (Eval)
MMStar (Multimodal)83.3 (Eval)
NL2Repo (Agents)37.9 (Eval)
NOVA-63 (General)57.9 (Eval)
ODinW (Multimodal)51.8 (Eval)
OmniDocBench 1.5 (Multimodal)91.2 (Eval)
OSWorld-Verified (Multimodal)62.5 (Eval)
PolyMATH (Math)77.4 (Eval)
RealWorldQA (Spatial Reasoning)85.4 (Eval)
RefCOCO-avg (Spatial Reasoning)0.9 (Eval)
ScreenSpot Pro (Multimodal)68.2 (Eval)
SimpleVQA (Image To Text)0.7 (Eval)
SkillsBench (Agents)45.7 (Eval)
SuperGPQA (Legal)71.6 (Eval)
SWE-bench Multilingual (Reasoning)73.8 (Eval)
SWE-Bench Pro (Reasoning)56.6 (Eval)
SWE-Bench Verified (Reasoning)78.8 (Eval)
TAU3-Bench (Reasoning)70.7 (Eval)
Terminal-Bench 2.0 (Reasoning)61.6 (Eval)
TIR-Bench (Multimodal)61.6 (Eval)
Toolathlon (Reasoning)39.8 (Eval)
V* (Multimodal)96.9 (Eval)
Video-MME (Multimodal)84.2 (Eval)
VideoMMMU (Multimodal)84.0 (Eval)
VITA-Bench (Reasoning)44.3 (Eval)
We-Math (Math)89.0 (Eval)
WideSearch (Reasoning)74.3 (Eval)
WMT24++ (Language)84.3 (Eval)
Artificial Analysis Intelligence Index (Overall)27.01 (Eval)
GPQA Diamond (Reasoning & Knowledge)0.9 (Eval)
Humanity's Last Exam (Reasoning & Knowledge)0.3 (Eval)
CritPt (Reasoning & Knowledge)0.0 (Eval)
AA-LCR (Long Context)0.8 (Eval)
IFBench (Instruction Following)0.8 (Eval)
τ²-Bench (Agentic)1.0 (Eval)
τ²-Bench Banking (Agentic)0.2 (Eval)
Terminal-Bench 2.1 (Agentic)0.6 (Eval)
Terminal-Bench Hard (Agentic)0.4 (Eval)
MMMU Pro (Multimodal)0.8 (Eval)
GDPval (Professional Tasks)993.01 (Eval)

API request example

Endpoint /v1/chat/completions. Replace <API_KEY> with a key from the console.

Standard OpenAI Chat Completions. Swap `model` for the target slug.

curl https://api.ravenrouter.ai/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer <API_KEY>" \
  -d '{
  "model": "Qwen3.6-Plus-2026-04-02",
  "messages": [
    {
      "role": "user",
      "content": "Hello! Introduce yourself in one sentence."
    }
  ],
  "max_tokens": 64
}'

More models: API Reference → Examples by model

Qwen3.6-Plus-2026-04-02 API — $1.29/1M input, $7.71/1M output | RavenRouter AI