跳轉到主要內容

Qwen3 Max Thinking

AlibabaQwenProprietary

描述

The latest flagship reasoning model in the Qwen3 family. Further enhanced by multiple innovations like adaptive tool-use and advanced test-time scaling techniques

發布日期
2026-01-26
參數規模
1.0T
上下文長度
262K
支援模態
text

能力雷達圖

60
general
50
coding
70
reasoning
77
science估算
40
agents
0
multimodal

缺少專門科學評測時,Science 由 LLM Stats 科學得分或推理能力估算。

排行榜排名

領域#排名分數來源
智慧體能力模型榜107
41.0
LS

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Agents

VITA-Bench40.9%自報
MCP-Mark33.5%自報
DeepPlanning28.7%自報
Toolathlon18.8%自報

Chat

IFEvalGoogle Research (2023)93.4%自報
Multi-Challenge63.3%自報

Code

SecCodeBench57.5%自報

General

C-Eval93.7%自報
MAXIFE84.0%自報
NOVA-6354.2%自報

Instruction Following

IFBench70.9%自報

Knowledge

HLE-Verified37.6%自報

Language

MMLU-Redux92.8%自報
MMLU-Pro85.7%自報
MMMLU84.4%自報
WMT24++77.6%自報

Long Context

LongBench v260.6%自報

Math

AIME 202693.3%自報
IMO-AnswerBench83.9%自報

Reasoning

GPQANYU + Cohere + Anthropic (2023)87.4%自報
Global PIQA86.0%自報
LiveCodeBench v685.9%自報
SWE-Bench Verified75.3%自報
AA-LCR68.7%自報
SuperGPQA67.3%自報
SWE-bench Multilingual66.7%自報
BrowseComp-zh60.9%自報
BrowseCompOpenAI (2025)53.9%自報
Humanity's Last Exam (with tools, text-only)49.8%自報
Seal-046.9%自報
Humanity's Last Exam (no tools, text-only)30.2%自報
Terminal-Bench 2.0Stanford × Laude Institute (2026)22.5%自報

Search

WideSearch57.9%自報

Tool Calling

BFCL-V467.7%自報

AA 評測指數

(Artificial Analysis)

暫無 AA 評測資料

LLM Stats 分類評分

(LLM Stats (zeroeval))
Language
90
Biology
90
Chat
80
Instruction Following
80
Legal
80
Physics
80
Structured Output
80
Finance
80
Frontend Development
80
Healthcare
80
Chemistry
80
Math
70
Economics
70
Long Context
60
Reasoning
60
General
60
Communication
60
Search
50
Code
50
Knowledge
40
Agents
40
Tool Calling
40

定價

輸入價格$1.2 / 1M tokens
輸出價格$6 / 1M tokens
混合價格(3:1)$2.4 / 1M tokens

速度

暫無速度資料

供應商價格排行

供應商價格排行

2 個供應商

最便宜: Alibaba最貴: DeepInfra
供應商輸入輸出
1Alibaba主要
$1.2
$6
2DeepInfra
$1.2
$6

比較該模型在不同 API 供應商之間的定價。

外部連結