Qwen3

Model mdl-9h4c6v

Back to Models
TU

Qwen3

Ready

Next-generation multimodal model unifying language and vision understanding

LanguageImage
Created

Mar 12, 2026

Parameters

32B

Accuracy

93.6%

Training Time

11h 20m

Data Sources

Production Corpus

s3://trinidy-training/corpus-v3

2.5M records

Image Dataset Alpha

gs://trinidy-media/images-alpha

890.0K records

Azure Telemetry Store

https://trinidydata.blob.core.windows.net/telemetry

5.1M records

Deploy this model to get an inference endpoint