메인 콘텐츠로 건너뛰기

Qwen3 VL 4B (Reasoning)

AlibabaQwen오픈 웨이트Apache 2.0 · 상업적 사용 가능

설명

Qwen3-VL is a large multimodal model that unifies vision, language, and reasoning to achieve human-level perception and cognition across text, images, and video. Built on a 235B-parameter architecture, it integrates early joint training of visual and textual modalities for strong language grounding. The model supports up to a 1 million-token context window and excels at visual understanding, spatial reasoning, long video comprehension, and tool-based interaction. It can generate code from images, perform precise 2D/3D object grounding, and operate digital interfaces like a visual agent. The “Instruct” version rivals Gemini 2.5 Pro in perception benchmarks, while the “Thinking” version leads in multimodal reasoning and STEM tasks. With multilingual OCR, creative writing, and fine-grained scene interpretation, Qwen3-VL establishes a new open-source frontier for integrated vision-language intelligence.

출시일
2025-10-14
파라미터
4.0B
컨텍스트 길이
—
모달리티
image, text

능력 레이더

26
general
32
coding
30
reasoning
36
science
70
agents
100
multimodal

랭킹

도메인#순위점수소스
코딩 랭킹506
20.0
AA
종합 랭킹498
28.0
AA
멀티모달 랭킹140
25.0
LS
과학513
26.0
AA

벤치마크 점수 (LLM Stats)

(LLM Stats (zeroeval))

Chat

MM-MT-Bench7.70 / 100자체 보고
IFEvalGoogle Research (2023)82.6%자체 보고
Multi-IF73.6%자체 보고

Creativity

Creative Writing v376.1%자체 보고

General

MMLU81.5%자체 보고
MLVU-M75.7%자체 보고
BFCL-v367.3%자체 보고
Include64.6%자체 보고
Arena-Hard v236.8%자체 보고

Language

MMLU-Redux86.0%자체 보고
MMLU-Pro73.6%자체 보고
MMLU-ProX65.0%자체 보고

Math

MathVista-Mini79.5%자체 보고
AIME 202574.5%자체 보고
LiveBench 2024112568.4%자체 보고
MathVision60.0%자체 보고
HMMT2553.1%자체 보고
PolyMATH44.6%자체 보고

Multimodal

VideoMMMU69.4%자체 보고
OSWorld31.4%자체 보고

Reasoning

CharXiv-D83.9%자체 보고
GPQANYU + Cohere + Anthropic (2023)64.1%자체 보고
LiveCodeBench v651.3%자체 보고
CharXiv-R50.3%자체 보고
SuperGPQA46.8%자체 보고

Video

CharadesSTA59.0%자체 보고

Vision

DocVQAtest94.2%자체 보고
ScreenSpot92.9%자체 보고
MMBench-V1.186.7%자체 보고
AI2D84.9%자체 보고
InfoVQAtest83.0%자체 보고
OCRBench80.8%자체 보고
MuirBench75.0%자체 보고
CC-OCR73.8%자체 보고
MMStar73.2%자체 보고
RealWorldQA73.2%자체 보고
MMMU (val)70.8%자체 보고
MVBench69.3%자체 보고
Hallusion Bench64.1%자체 보고
BLINK63.4%자체 보고
OCRBench-V2 (en)61.8%자체 보고
MMMU-Pro57.0%자체 보고
OCRBench-V2 (zh)55.8%자체 보고
LVBench53.5%자체 보고
ScreenSpot Pro49.2%자체 보고
ERQA47.3%자체 보고
ODinW39.4%자체 보고

Writing

WritingBench84.0%자체 보고

AA 평가 지수

(Artificial Analysis)
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))
70.0
Gpqa(NYU + Cohere + Anthropic (2023))
49.4
Ifbench(Google Research (2023))
36.6
Livecodebench(UC Berkeley + MIT + Cornell (2024))
32.0
Math Index(Artificial Analysis)
25.7
Aime 25(MAA (Mathematical Association of America))
25.7
Lcr(Artificial Analysis)
21.3
Tau2(Sierra + U Toronto + Vector Institute (2025))
15.5
Intelligence Index(Artificial Analysis)
7.0
Hle(Center for AI Safety + Scale AI (2025))
4.6
Terminalbench Hard(Stanford × Laude Institute (2026))
1.5

LLM Stats 카테고리 점수

(LLM Stats (zeroeval))
Chat
3
Communication
3
Multimodal
100
Instruction Following
80
Structured Output
80
General
80
Image To Text
70
Language
70
Legal
70
Math
70
Reasoning
70
Finance
70
Grounding
70
Healthcare
70
Creativity
70
Text-to-image
70
Tool Calling
70
Vision
70
Writing
70
Physics
60
Spatial Reasoning
60
3d
60
Biology
60
Chemistry
60
Video
60
Long Context
50
Agents
50
Economics
50

가격

입력 가격무료
출력 가격무료
혼합 가격 (3:1)무료

속도

토큰/초0.0
첫 토큰 지연0.00s
첫 응답 지연0.00s

공급자 가격 순위

프로바이더 데이터가 없습니다

외부 링크