Feed / Azure
mediumAzureazure-ai-evaluationupdategithub_azure_sdk_for_python · Jul 22, 2026

azure-ai-evaluation 1.18.2

The azure-ai-evaluation SDK received an update with support for admin-connected (BYO) judge models in prompty-based evaluators, allowing models configured with 'byo_model' and 'project_endpoint' to be routed through the Foundry project Responses API. Additionally, validation errors in TaskNavigationEfficiencyEvaluator were fixed.

View original article →
Audience
developer
Environment
cloud
Classification confidence: 95%
Loading correlations…
azure-ai-evaluation 1.18.2 | 365Forge