Qwen: Qwen3 VL 32B Instruct
qwen/qwen3-vl-32b-instruct
150th smartest of 182Cheaper than 82% of paidToolsJSONVision
Intelligence
11.1
150th of 182
Design Elo
—
Speed
—
tokens/sec
Latency
—
first token
Input price
$0.104
60th cheapest
Context
131K
33K max out
How it compares
Smarter than18%
of all ranked models
Cheaper than82%
of all ranked models
Overview
Qwen3-VL-32B-Instruct is a large-scale multimodal vision-language model designed for high-precision understanding and reasoning across text, images, and video. With 32 billion parameters, it combines deep visual perception with advanced text...
Benchmarks
independent · Artificial Analysis & Design ArenaArtificial Analysis39th percentile
Intelligence Index
11.1
GPQA Diamond
67%
Humanity's Last Exam
6%
SciCode
30%
Tau²-Bench (agentic)
29%
Providers & pricing (1)
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| Alibabafp8 | $0.104 | $0.416 | 131K | 100% |
Specifications
Context window131K
Max output33K
Knowledge cutoff—
Input modalitiestext, image
Output modalitiestext
Prompt caching—
Cache read price—
ModeratedNo
Open weightsQwen/Qwen3-VL-32B-Instruct ↗
Qwen3 VL 32B Instruct FAQ
How much does Qwen3 VL 32B Instruct cost?
Qwen3 VL 32B Instruct costs $0.104 per million input tokens and $0.416 per million output tokens via OpenRouter, making it 60th cheapest of 332 paid models.
How smart is Qwen3 VL 32B Instruct?
Qwen3 VL 32B Instruct scores 11.1 on the Artificial Analysis Intelligence Index, ranking 150th of 182 benchmarked models, with a GPQA Diamond score of 67%.
What is Qwen3 VL 32B Instruct's context window?
Qwen3 VL 32B Instruct supports a 131K-token context window and can output up to 33K tokens. It accepts text, image input.
Similar models
All Qwen3 VL 32B Instruct alternatives →Compare head-to-head
Qwen: Qwen3 VL 32B Instruct vs Z.ai: GLM 4.6VQwen: Qwen3 VL 32B Instruct vs OpenAI: GPT-4o (2024-11-20)Qwen: Qwen3 VL 32B Instruct vs OpenAI: GPT-4oQwen: Qwen3 VL 32B Instruct vs Qwen: Qwen3 VL 30B A3B InstructQwen: Qwen3 VL 32B Instruct vs Meta: Llama 4 ScoutQwen: Qwen3 VL 32B Instruct vs DeepSeek: R1 Distill Llama 70B