跳转到主要内容

Qwen3.5 4B (Non-reasoning)

AlibabaQwen开源权重Apache 2.0 · 商用许可

描述

Qwen3.5-4B is a 4 billion parameter vision-language model using Gated DeltaNet hybrid architecture with a 3:1 ratio of linear attention to full softmax attention. It supports 262K native context length and delivers strong performance for its size across knowledge, reasoning, coding, and multilingual tasks.

发布日期
2026-03-02
参数规模
4.0B
上下文长度
支持模态

能力雷达图

14
general
20
coding
71
reasoning
40
science估算
70
agents
50
multimodal

Science 在缺少专门科学评测时使用推理能力代理估算。

排行榜排名

领域#排名分数来源
智能体能力模型榜95
47.0
LS
代码能力榜310
27.0
AA
通用能力榜248
42.0
AA
科学能力304
38.0
AA

基准测试分数 (LLM Stats)

Agents

t2-bench79.9%自报
BFCL-V450.3%自报
VITA-Bench22.0%自报
DeepPlanning17.6%自报

Biology

GPQA76.2%自报

Chemistry

SuperGPQA52.9%自报

Communication

Multi-Challenge49.0%自报

Finance

MMLU-Pro79.1%自报
MMLU-ProX71.5%自报

General

IFEval89.8%自报
MMLU-Redux88.8%自报
C-Eval85.1%自报
Global PIQA78.9%自报
MAXIFE78.0%自报
MMMLU76.1%自报
Include71.0%自报
IFBench59.2%自报
LiveCodeBench v655.8%自报
NOVA-6354.3%自报
LongBench v250.0%自报

Language

WMT24++66.6%自报

Long Context

AA-LCR57.0%自报

Math

HMMT2576.8%自报
HMMT 202574.0%自报
PolyMATH51.1%自报

AA 评测指数

Coding Index
20.3
Intelligence Index
16.0
Tau2
0.9
Gpqa
0.7
Ifbench
0.3
Lcr
0.3
Terminalbench V2 1
0.2
Scicode
0.2
Terminalbench Hard
0.1
Hle
0.1
Tau Banking
0.0

LLM Stats 分类评分

Language
80
Biology
80
Instruction Following
70
Legal
70
Math
70
Physics
70
Structured Output
70
Finance
70
General
70
Healthcare
70
Tool Calling
70
Reasoning
60
Chemistry
60
Long Context
50
Multimodal
50
Spatial Reasoning
50
Communication
50
Economics
50
Vision
50
Agents
40

定价

输入价格$0.03 / 1M tokens
输出价格$0.15 / 1M tokens
混合价格(3:1)$0.06 / 1M tokens

速度

Tokens/秒40.6
首Token延迟0.43s
首回答延迟0.43s

供应商价格排行

供应商价格排行

1 个供应商

供应商输入输出
1Alibaba主要
$0.03
$0.15

比较该模型在不同 API 供应商之间的定价。

外部链接