Add support for GPT-5 model family (#2667)

Author: pamelafox

app/backend/approaches/approach.py

@@ -136,6 +136,9 @@ class Approach(ABC):
         "o3": GPTReasoningModelSupport(streaming=True),
         "o3-mini": GPTReasoningModelSupport(streaming=True),
         "o4-mini": GPTReasoningModelSupport(streaming=True),
+        "gpt-5": GPTReasoningModelSupport(streaming=True),
+        "gpt-5-nano": GPTReasoningModelSupport(streaming=True),
+        "gpt-5-mini": GPTReasoningModelSupport(streaming=True),
     }
     # Set a higher token limit for GPT reasoning models
     RESPONSE_DEFAULT_TOKEN_LIMIT = 1024

docs/deploy_features.md

@@ -35,10 +35,16 @@ As of early June 2025, the default chat completion model is `gpt-4.1-mini`. If y
     For example:
 
     ```bash
-    azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT gpt-4o
+    azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT gpt-5-chat
     ```
 
-1. To set the GPT model to a different [available model](https://learn.microsoft.com/azure/ai-services/openai/concepts/models), run this command with the appropriate model name.
+1. To set the GPT model to a different [available model](https://learn.microsoft.com/azure/ai-services/openai/concepts/models), run this command with the appropriate model name. For reasoning models like gpt-5/o3/o4, check [the reasoning guide](./reasoning.md)
+
+   For gpt-5-chat:
+
+   ```shell
+   azd env set AZURE_OPENAI_CHATGPT_MODEL gpt-5-chat
+   ```
 
     For gpt-4.1-mini:
 
@@ -72,6 +78,12 @@ As of early June 2025, the default chat completion model is `gpt-4.1-mini`. If y
 
 1. To set the Azure OpenAI model version from the [available versions](https://learn.microsoft.com/azure/ai-services/openai/concepts/models), run this command with the appropriate version string.
 
+   For gpt-5-chat:
+
+   ```shell
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_VERSION 2025-08-07
+   ```
+
     For gpt-4.1-mini:
 
     ```bash

docs/reasoning.md

@@ -6,6 +6,9 @@ This repository includes an optional feature that uses reasoning models to gener
 
 ### Supported Models
 
+* gpt-5
+* gpt-5-mini
+* gpt-5-nano
 * o4-mini
 * o3
 * o3-mini
@@ -21,6 +24,36 @@ This repository includes an optional feature that uses reasoning models to gener
 
    Set the environment variables for your Azure OpenAI GPT deployments to your reasoning model
 
+   For gpt-5:
+
+   ```shell
+   azd env set AZURE_OPENAI_CHATGPT_MODEL gpt-5
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT gpt-5
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_VERSION 2025-08-07
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_SKU GlobalStandard
+   azd env set AZURE_OPENAI_API_VERSION 2025-04-01-preview
+   ```
+
+   For gpt-5-mini:
+
+   ```shell
+   azd env set AZURE_OPENAI_CHATGPT_MODEL gpt-5-mini
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT gpt-5-mini
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_VERSION 2025-08-07
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_SKU GlobalStandard
+   azd env set AZURE_OPENAI_API_VERSION 2025-04-01-preview
+   ```
+
+   For gpt-5-nano:
+
+   ```shell
+   azd env set AZURE_OPENAI_CHATGPT_MODEL gpt-5-nano
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT gpt-5-nano
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_VERSION 2025-08-07
+   azd env set AZURE_OPENAI_CHATGPT_DEPLOYMENT_SKU GlobalStandard
+   azd env set AZURE_OPENAI_API_VERSION 2025-04-01-preview
+   ```
+
    For o4-mini:
 
    ```shell
@@ -63,12 +96,12 @@ This repository includes an optional feature that uses reasoning models to gener
 
 2. **(Optional) Set default reasoning effort**
 
-   You can configure how much effort the reasoning model spends on processing and understanding the user's request. Valid options are `low`, `medium`, and `high`. Reasoning effort defaults to `medium` if not set.
+   You can configure how much effort the reasoning model spends on processing and understanding the user's request. Valid options are `minimal` (for GPT-5 models only), `low`, `medium`, and `high`. Reasoning effort defaults to `medium` if not set.
 
-   Set the environment variable for reasoning effort
+   Set the environment variable for reasoning effort:
 
    ```shell
-   azd env set AZURE_OPENAI_REASONING_EFFORT medium
+   azd env set AZURE_OPENAI_REASONING_EFFORT minimal
    ```
 
 3. **Update the infrastructure and application:**