Qwen: Qwen3 VL 30B A3B Thinking

Qwen: Qwen3 VL 30B A3B Thinking

qwen · Released Oct 6, 2025 Efficient
49
Our Score

Performance Profile

Intelligence4.2Technical2.7Value7.8Content4.5
Intelligence 4.2/10
Technical 2.7/10
Content 4.5/10
Value 7.8/10

Qwen3-VL-30B-A3B-Thinking is a multimodal model that unifies strong text generation with visual understanding for images and videos. Its Thinking variant enhances reasoning in STEM, math, and complex tasks. It excels..

$0.13 / 1M
Input Price
$1.56 / 1M
Output Price
131,072 tokens
Context Window
32,768 tokens
Max Output
30B Parameters

Capabilities

Tool Use Function Calling Vision

Architecture

ModalityText + Image → Text
TokenizerQwen3
Parameters30B

Performance Indices

Source: Artificial Analysis

19.7 Intelligence Index
13.1 Coding Index
12.6 Agentic Index
82.3 Math Index

Benchmark Scores

Intelligence

GPQA Diamond 72% Graduate-level scientific reasoning
HLE 8.7% Humanity's Last Exam
MMLU Pro 80.7% Multi-task language understanding
AIME 2025 82.3% Competition mathematics (2025)
SciCode 28.8% Scientific computing

Technical

LiveCodeBench 69.7% Live coding evaluation
TerminalBench Hard 5.3% Agentic terminal tasks
τ²-Bench 19.9% Conversational agent benchmark

Content

IFBench 45.1% Instruction following
LCR 40.7% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does Qwen: Qwen3 VL 30B A3B Thinking stack up?

Compare side-by-side with other efficient models.

Compare Models

Model Information

OpenRouter ID qwen/qwen3-vl-30b-a3b-thinking
Providerqwen
Release Date October 6, 2025
Context Length131,072 tokens
Max Completion32,768 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.13 $0.000130
Output $1.56 $0.001560

Live Performance

Live endpoint metrics — refreshed every 30 minutes.

99.8%
Avg Uptime
968ms
Best Latency (TTFT)
121 tok/s
Best Throughput
2/3
Active Endpoints
Available via: Alibaba, Novita, SiliconFlow