Fix GPT 5.6+ Multi-Modal Vision model selection

[paullizer]

Summary

• recognize GPT 5.6 and later vision-capable aliases across model, display, and deployment names
• keep initial Jinja rendering and client-side endpoint updates on the same capability contract
• preserve endpoint/model enablement filtering and improve labels when providers omit model names
• bump the application to v0.250.066 and update fix, feature, and release documentation

Validation

• python -m pytest ui_tests/test_admin_multimodal_vision_model_options.py functional_tests/test_multimodal_vision_multi_endpoint_connection.py functional_tests/test_vision_model_parameter_fix.py functional_tests/route_tests/test_route_blueprint_policy_inventory.py functional_tests/route_tests/test_route_unauthenticated_policy_contract.py functional_tests/route_tests/test_route_policy_test_coverage.py -q (17 passed)
• node --check application/single_app/static/js/admin/admin_settings.js
• python -m py_compile application/single_app/functions_model_capabilities.py application/single_app/route_frontend_admin_settings.py ui_tests/test_admin_multimodal_vision_model_options.py
• git diff --check

Fixes #1086