Azure OpenAI GPT-6.1 Sol – Significant latency degradation across Spain and Sweden Data Zones

Hello,

We are experiencing a significant increase in response latency when using Azure OpenAI GPT-6.1 Sol deployments across two different Data Zones: Spain and Sweden.

The degradation started during the week of November 9–13, 2026.

Environment:

• Service: Azure OpenAI

Model: GPT-6.1 Sol

Deployment type: Data Zone Standard

Data Zones: Spain and Sweden

Deployments: gpt-54-bbva and gpt-54

Production workload: High-volume conversational AI and real-time voice agents.

Important details:

We have not made any changes to our infrastructure or application during this period:

No backend code changes.

No API request parameter changes.

No prompt changes.

No model configuration changes.

No intentional changes to traffic patterns.

Both Data Zones are experiencing increased latency simultaneously.

Questions:

Are there any known performance degradation issues affecting GPT-6.1 Sol inference during November 9–13, 2026?

Have there been any recent changes to Azure OpenAI inference infrastructure, GPU allocation, request routing or load balancing?

Could a shared infrastructure component explain simultaneous latency degradation across Spain and Sweden Data Zones?

Is Microsoft aware of any capacity constraints or performance regressions affecting GPT-6.1 Sol?

This issue is affecting our production environment, where low response latency is critical for real-time voice conversations.

We would appreciate any information from Microsoft engineers regarding potential infrastructure issues or ongoing investigations.

Thank you.Hello,

We are experiencing a significant increase in response latency when using Azure OpenAI GPT-6.1 Sol deployments across two different Data Zones: Spain and Sweden.

The degradation started during the week of November 9–13, 2026.

Environment:

• Service: Azure OpenAI
• Model: GPT-6.1 Sol
• Deployment type: Data Zone Standard
• Data Zones: Spain and Sweden
• Production workload: High-volume conversational AI and real-time voice agents.

Important details:

We have not made any changes to our infrastructure or application during this period:

No backend code changes.

No API request parameter changes.

No prompt changes.

No model configuration changes.

No intentional changes to traffic patterns.

Both Data Zones are experiencing increased latency simultaneously.

Questions:

Are there any known performance degradation issues affecting GPT-6.1 Sol inference during November 9–13, 2026?

Have there been any recent changes to Azure OpenAI inference infrastructure, GPU allocation, request routing or load balancing?

Could a shared infrastructure component explain simultaneous latency degradation across Spain and Sweden Data Zones?

Is Microsoft aware of any capacity constraints or performance regressions affecting GPT-6.1 Sol?

This issue is affecting our production environment, where low response latency is critical for real-time voice conversations.

We would appreciate any information from Microsoft engineers regarding potential infrastructure issues or ongoing investigations.

Thank you.