Qwen: Qwen3 VL 30B A3B Thinking

Image input Text input Text output
Author's Description

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 in perception of real-world/synthetic categories, 2D/3D spatial grounding, and long-form visual comprehension, achieving competitive multimodal benchmark results. For agentic use, it handles multi-image multi-turn instructions, video timeline alignments, GUI automation, and visual coding from sketches to debugged UI. Text performance matches flagship Qwen3 models, suiting document AI, OCR, UI assistance, spatial tasks, and agent research.

Key Specifications
Cost
$$$$
Context
262K
Parameters
30B
Released
Oct 06, 2025
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Stop Reasoning Logit Bias Seed Presence Penalty Temperature Max Tokens Top P Structured Outputs Min P Tool Choice Frequency Penalty Tools Include Reasoning Response Format
Features

This model supports the following features:

Structured Outputs Tools Reasoning Response Format
Performance Summary

Qwen: Qwen3 VL 30B A3B Thinking demonstrates moderate speed performance, ranking in the 23rd percentile across benchmarks, and offers moderate pricing, placing it in the 40th percentile. Notably, it exhibits exceptional reliability with a 99% success rate, indicating minimal technical failures. The model excels in ethical reasoning, achieving perfect accuracy and being the most accurate at its price point and speed. It also shows strong performance in Email Classification (99.0% accuracy) and Mathematics (91.9% accuracy). Its General Knowledge is solid at 97.0% accuracy. However, a significant weakness is observed in Hallucinations, where it only achieved 84.0% accuracy, indicating a tendency to not acknowledge uncertainty. Reasoning performance is moderate at 67.3% accuracy. This model is particularly well-suited for applications requiring high reliability and strong ethical considerations, alongside robust mathematical and classification capabilities, despite its moderate speed and a tendency to hallucinate on fictional concepts.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.16
Completion $0.8

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
Parasail
Parasail | qwen/qwen3-vl-30b-a3b-thinking 262K $0.16 / 1M tokens $0.8 / 1M tokens
SiliconFlow
SiliconFlow | qwen/qwen3-vl-30b-a3b-thinking 262K $0.29 / 1M tokens $1 / 1M tokens
Novita
Novita | qwen/qwen3-vl-30b-a3b-thinking 131K $0.16 / 1M tokens $0.8 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by qwen