Qwen3 VL 4B Instruct is Alibaba's language model. A compact 4B instruction-tuned vision-language model from the Qwen3 series, balancing multimodal understanding with efficient inference.
Specifications
Canonical IDalibaba-qwen3-vl-4b-instruct
TypeLanguage
StatusActive
CreatorAlibabaAlibaba
Input ModalitiesText
Output ModalitiesText
Parameters4B
Benchmarks
Intelligence Index
9.6
#396
Coding Index
4.6
#343
Math Index
37.0
#159
MMLU-Pro
0.6
#239
GPQA
0.4
#371
HLE
0.0
#405
LiveCodeBench
0.3
#214
IFBench
0.3
#305
Time to First Token
SciCode
0.1
#369
AIME 2025
0.4
#159
LCR
0.1
#273
TerminalBench Hard
0.0
#335
TAU2
0.2
#263
Output TPS
0.0
#277

Capabilities

Input1/5
Text
Image·
Audio·
Video·
PDF·
Output1/5
Text
Image·
Audio·
Video·
Embedding·
Capabilities0/13
Reasoning·
Adaptive Reasoning·
Function Calling·
Parallel Function Calling·
Structured Outputs·
Native JSON Schema·
Web Search·
URL Context·
Computer Use·
Code Execution·
File Search·
Prompt Caching·
Assistant Prefill·

Versions

VersionReleasedContextInput / 1MOutput / 1MStatus
EAGLE Qwen 2.5 3B InstructAvailable
Qwen3.6 Max Preview262K$1.04$6.24Available
Qwen3.6 27B262K$0.320$3.20Available
Qwen3.6 35B A3B262K$0.150$1.00Available
Qwen3.6 Plus1.0M$0.325$1.95Available
Qwen3 Max Thinking262K$0.780$3.90Available
Qwen3 Next 80B A3B128K$0.140$1.20Available
Qwen3 Max262K$0.359$1.43Available
Qwen3 Max Preview262K$1.20$6.00Available
Qwen3 Coder Plus1.0M$0.650$3.25Available
Qwen3 VL 4B InstructCurrent

Model IDs