跳转到主要内容
AITier

Qwen3.5 397B A17B (Non-reasoning)

AlibabaQwen开源权重Apache 2.0 · 商用许可

描述

Qwen3.5-397B-A17B is Qwen's flagship Mixture-of-Experts model with 397 billion total parameters and 17 billion activated parameters. It delivers state-of-the-art performance across knowledge, reasoning, coding, mathematics, multilingual understanding, instruction following, long context, and agent tasks.

发布日期
2026-02-16
参数规模
397.0B
上下文长度
262K
支持模态
audio, image, text, video

能力雷达图

21
general
70
coding
86
reasoning
66
science
60
agents
70
multimodal

排行榜排名

领域#排名分数来源
智能体能力模型榜17
62.0
LS
代码能力榜210
64.0
AA
通用能力榜207
52.0
AA
科学能力189
60.0
AA

基准测试分数 (LLM Stats)

(LLM Stats (zeroeval))

Agents

t2-bench86.7%自报
VITA-Bench49.7%自报
MCP-Mark46.1%自报
Toolathlon38.3%自报
DeepPlanning34.3%自报

Chat

IFEvalGoogle Research (2023)92.6%自报
Multi-Challenge67.6%自报

Code

SecCodeBench68.3%自报

General

C-Eval93.0%自报
MAXIFE88.2%自报
Include85.6%自报
NOVA-6359.1%自报

Instruction Following

IFBench76.5%自报

Language

MMLU-Redux94.9%自报
MMMLU88.5%自报
MMLU-Pro87.8%自报
MMLU-ProX84.7%自报
WMT24++78.9%自报

Long Context

LongBench v263.2%自报

Math

HMMT 202594.8%自报
HMMT2592.7%自报
AIME 202691.3%自报
IMO-AnswerBench80.9%自报
PolyMATH73.3%自报

Reasoning

Global PIQA89.8%自报
GPQANYU + Cohere + Anthropic (2023)88.4%自报
LiveCodeBench v683.6%自报
SWE-Bench Verified76.4%自报
SuperGPQA70.4%自报
BrowseComp-zh70.3%自报
SWE-bench Multilingual69.3%自报
BrowseCompOpenAI (2025)69.0%自报
AA-LCR68.7%自报
Terminal-Bench 2.0Stanford × Laude Institute (2026)52.5%自报
Seal-046.9%自报
Humanity's Last Exam28.7%自报

Search

WideSearch74.0%自报

Tool Calling

BFCL-V472.9%自报

AA 评测指数

(Artificial Analysis)
Gpqa(NYU + Cohere + Anthropic (2023))
86.1
Tau2(Sierra + U Toronto + Vector Institute (2025))
83.9
Lcr(Artificial Analysis)
64.3
Ifbench(Google Research (2023))
51.6
Terminalbench Hard(Stanford × Laude Institute (2026))
35.6
Intelligence Index(Artificial Analysis)
21.4
Hle(Center for AI Safety + Scale AI (2025))
19.8

LLM Stats 分类评分

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

定价

输入价格$0.6 / 1M tokens
输出价格$3.6 / 1M tokens
混合价格(3:1)$1.35 / 1M tokens

速度

Tokens/秒86.3
首Token延迟1.70s
首回答延迟1.70s

供应商价格排行

供应商价格排行

2 个供应商

最便宜: DeepInfra最贵: Alibaba
供应商输入输出
1DeepInfra最便宜
$0
$0
2Alibaba主要
$0.6
$3.6

比较该模型在不同 API 供应商之间的定价。

外部链接

LLM StatsArtificial Analysis