Microsoft Adds GPT-5.6, GPT-6 and GPT-6.1 to Azure Content Understanding

Microsoft adds GPT-5.6, GPT-6 and GPT-6.1 support to Azure Content Understanding, detailed in a company blog post. The models are available for document, text, speech and video workflows.

Microsoft recommends GPT-6.1 Sol as the starting option for document extraction. Its evaluation found that GPT-6 Astra recorded 95.95% accuracy with a schema alone, compared with 95.74% for Sol, but Astra used about five times as much in completion token costs. GPT-6 Luna offered lower token costs, while five labeled examples increased its extraction score from 91.82% to 94.16%.

Results varied by task. GPT-6 Astra led Microsoft's document segmentation benchmark, GPT-6 Sol posted the highest overall video generation score, and GPT-6.1 Sol performed best on short video ads. Customers can compare model deployments in Content Understanding Studio or through the REST API while keeping the analyzer, schema and input files unchanged.