Intelligence Index
0111.0
> 16% OF MODELS
Qwen
Qwen3 VL 32B Instruct
Updated: September 2026
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...
Specifications
- Context
- 131K
- Input Modalities
- Text, Image
- Output Modalities
- Text
- Input
- $0.104/M
- Output
- $0.416/M
Capabilities
VISIONTEXTWEBTHINKINGWRITING
Similarly Priced Models
| Model | Provider | Context | Input Price | Output Price |
|---|---|---|---|---|
| GPT-5.6 Luna | OpenAI | 1M | $0.1/M | $0.6/M |
| GPT-5.6 Luna | OpenAI | 1M | $0.1/M | $0.6/M |
| GPT-5.6 Luna | OpenAI | 1M | $0.1/M | $0.6/M |
| GPT-5.6 Luna | OpenAI | 1M | $0.1/M | $0.6/M |
| Qwen3.5-9B | Qwen | 262K | $0.1/M | $0.15/M |
Performance Metrics
Average Response Performance
- Output Speed
- 56.3 tok/s
- Time To First Token
- 2.68s
- Time To First Answer Token
- 2.68s
- End To End Response Time
- 11.56s
DATA SOURCE: Artificial Analysis