Supported OpenAI models for AI integration
Introduced in 2024.2
Updated in 2026
This article explains which AI models you need to configure for the AI integration in Alloy Navigator and what each model is used for. You select models during AI integration, when entering your API key and configuring AI features.
The AI integration uses models from OpenAI. You can connect using one of the following AI providers:
-
OpenAI API - provides a quick and flexible setup
-
Azure OpenAI Service - offers additional enterprise capabilities such as regional deployment, compliance options, and integration with Microsoft services
IMPORTANT: If you use Azure OpenAI Service, a deployment for each model you plan to use must be created. These deployments are typically configured by your Azure administrator. To complete AI integration, you will need to specify these deployment names in Alloy Navigator.
Use the table below for a quick overview and to explore details as needed.
| Model | Required or optional | Description | AI feature usage |
|---|---|---|---|
|
Required |
Processes a prompt and produces a context-aware response |
||
|
Optional |
Converts written text into spoken words |
||
|
Optional |
Transcripts spoken language into text |
||
|
Optional, recommended |
Generates vector representations of text for semantic search and internal knowledge access |
INFO: For a full listing of available OpenAI models, see https://platform.openai.com/docs/models. For information about Azure OpenAI Service models, see https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models.
Completion model (AI response generation model)
The completion model is an AI model that generates responses by continuing a given input (prompt). Completion models support a wide range of tasks, such as text generation, question answering, and language translation. It is a core component of all AI-powered features in Alloy Navigator, including AI-Powered Insights, AI Writing Assistance, and AI Assistant for Self Service Portal.
By default, Alloy Navigator uses gpt-5.4-nano, which is optimized for everyday use and offers an effective balance between speed, cost, and quality.
We recommend the following models, listed from lightweight to more advanced:
-
gpt-5.4-nano (default) - balanced performance, cost, and quality
-
gpt-5-mini - cost-efficient option for simple or high-volume tasks
-
gpt-5.4-mini - improved response quality with moderate cost
-
gpt-5.4 -— suitable for complex tasks requiring higher accuracy
-
gpt-5.5 - highest capability currently available; higher cost compared to other models
NOTE: Mini and nano models are typically faster and more cost-efficient, but offer lower capability than larger models.
New AI models are released more frequently than Alloy Navigator updates, so the dropdown list in the AI Configuration section in Alloy Navigator may not always include the latest models. You can choose other models based on your needs (see the OpenAI documentation for available options) and enter any model available in your AI provider.
If you use Azure OpenAI Service, the same model recommendations apply. Make sure your deployment uses a GPT model that corresponds to one of the recommended completion models.
Text-to-Speech model
AI Writing Assistance uses a Text-to-Speech model to convert written text into natural-sounding speech.
We recommend the following Text-to-Speech models.
-
tts-1 (default) - best balance of performance, cost, and quality
-
tts-1-hd - higher-quality output with increased cost
If you use Azure OpenAI Service, the same model recommendations apply. Make sure your deployment uses an audio (Text-to-Speech) model.
NOTE: Using a Text-to-Speech model is optional. If you do not plan to use voice output, you can disable this feature. AI Writing Assistance will continue to work without it.
Transcription model
AI Writing Assistance uses a transcription model to convert speech into written text.
By default, Alloy Navigator uses whisper-1, a reliable and accurate model for speech-to-text transcription. You can also select or enter a different model supported by your AI provider.
If you use Azure OpenAI Service, make sure your deployment uses an audio (transcription) model.
NOTE: Using a transcription model is optional. If you do not plan to use speech recognition, you can disable this feature. AI Writing Assistance will continue to work without it.
Embedding model
The embedding model generates vector representations of text for semantic search and access to internal knowledge in for AI Assistant for Self Service Portal and AI-Powered Insights.
We recommend the following embedding models:
-
text-embedding-3-large (default) - higher quality embeddings, suitable for more accurate semantic search
-
text-embedding-3-small - more cost-efficient option with slightly lower accuracy
You can also select or enter a different model supported by your AI provider.
If you use Azure OpenAI Service, make sure your deployment uses an embedding model.
IMPORTANT: Using an embedding model is optional, but highly recommended. If you disable access to internal knowledge, AI Assistant and AI-Powered Insights will continue to work, but responses may be less accurate and relevant, as the system will not be able to use your Service Desk knowledge (ticket history and knowledge base).