跳轉到主要內容
AITier

Qwen3.5 397B A17B (Non-reasoning)

AlibabaQwen開源權重Apache 2.0 · 商用許可

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.

發布日期
2026-02-16
參數規模
397.0B
上下文長度
262K
支援模態
audio, image, text, video

能力雷達圖

29
general
41
coding
86
reasoning
57
science估算
60
agents
70
multimodal

Science 在缺少專門科學評測時使用推理能力代理估算。

排行榜排名

領域#排名分數來源
智慧體能力模型榜60
55.0
LS
程式碼能力榜100
67.0
AA
通用能力榜110
62.0
AA
推理能力69
58.0
LS
科學能力80
64.0
AA

基準測試分數 (LLM Stats)

Agents

GDPval-AA961.00 / 3000自報
t2-bench86.7%自報
WideSearch74.0%自報
BFCL-V472.9%自報
BrowseComp69.0%自報
Terminal-Bench 2.052.5%自報
VITA-Bench49.7%自報
MCP-Mark46.1%自報
Toolathlon38.3%自報
DeepPlanning34.3%自報

Biology

GPQA88.4%自報

Chemistry

SuperGPQA70.4%自報

Code

SWE-Bench Verified76.4%自報
SWE-bench Multilingual69.3%自報

Coding

SecCodeBench68.3%自報

Communication

Multi-Challenge67.6%自報

Finance

MMLU-Pro87.8%自報
MMLU-ProX84.7%自報

General

MMLU-Redux94.9%自報
C-Eval93.0%自報
IFEval92.6%自報
Global PIQA89.8%自報
MMMLU88.5%自報
MAXIFE88.2%自報
Include85.6%自報
LiveCodeBench v683.6%自報
IFBench76.5%自報
LongBench v263.2%自報
NOVA-6359.1%自報

Language

WMT24++78.9%自報

Long Context

AA-LCR68.7%自報

Math

HMMT 202594.8%自報
HMMT2592.7%自報
AIME 202691.3%自報
IMO-AnswerBench80.9%自報
PolyMATH73.3%自報
Humanity's Last Exam28.7%自報

Reasoning

BrowseComp-zh70.3%自報
Seal-046.9%自報

AA 評測指數

Intelligence Index
32.0
Gpqa
0.9
Tau2
0.8
Lcr
0.6
Ifbench
0.5
Scicode
0.4
Terminalbench Hard
0.4
Hle
0.2

LLM Stats 分類評分

Legal
100
Finance
100
Agents
97
General
61
Reasoning
35
Language
90
Biology
90
Instruction Following
80
Math
80
Physics
80
Structured Output
80
Frontend Development
80
Healthcare
80
Chemistry
80
Long Context
70
Multimodal
70
Search
70
Spatial Reasoning
70
Code
70
Communication
70
Economics
70
Tool Calling
60
Vision
50

定價

輸入價格$0.6 / 1M tokens
輸出價格$3.6 / 1M tokens
混合價格(3:1)$1.35 / 1M tokens

速度

Tokens/秒52.9
首Token延遲1.75s
首回答延遲1.75s

供應商價格排行

供應商價格排行

9 個供應商

最便宜: Novita最貴: OVHcloud AI Endpoints
供應商輸入輸出
1Novita最便宜
$0
$0
2CrofAI
$0.35
$1.75
3Alibaba (China)
$0.43
$2.58
4DigitalOcean
$0.55
$3.5
5Alibaba主要
$0.6
$3.6
6Scaleway
$0.6
$3.6
7NanoGPT
$0.6
$3.6
8Cortecs
$0.6
$3.6
9OVHcloud AI Endpoints
$0.71
$4.25

比較該模型在不同 API 供應商之間的定價。

外部連結

LLM StatsArtificial Analysis