MultimodalOpen weights

Qwen3 VL 32B Instruct

Alibaba

Updated May 22, 2026 · Benchmarks via Artificial Analysis, specs & pricing via OpenRouter · Methodology · Report an error

Specifications

Type: Multimodal
Access: Open weights
Released: October 23, 2025
Context window: 262,144 tokens
Output speed: 76 tok/s
Latency (TTFT): 1.16s
Input: Text, Image
Output: Text
API pricing: $0.10 in · $0.42 out / 1M tokens

Capabilities

Function callingStructured output

Benchmarks

Reasoning

67.1

Coding

51.4

Math

68.3

General

MMLU-Pro	79.1
Humanity’s Last Exam	6.3

Compare Qwen3 VL 32B Instruct with

vs Sonar Reasoning Pro vs R1 1776 vs Qwen3.7 Max vs Gemini 3.5 Flash

See all Qwen3 VL 32B Instruct alternatives →