You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Copy file name to clipboardExpand all lines: articles/ai-services/openai/concepts/gpt-4-v-prompt-engineering.md
+2-2Lines changed: 2 additions & 2 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -6,13 +6,13 @@ author: PatrickFarley
6
6
ms.author: pafarley
7
7
ms.service: azure-ai-openai
8
8
ms.topic: conceptual
9
-
ms.date: 02/20/2025
9
+
ms.date: 05/31/2025
10
10
manager: nitinme
11
11
---
12
12
13
13
# Image prompt engineering techniques
14
14
15
-
To unlock the full potential of vision-enabled chat models like GPT-4 Turbo with Vision and GPT-4o, it's essential to tailor the prompts to your specific needs. Here are some guidelines to enhance the accuracy and efficiency of your prompts.
15
+
To unlock the full potential of vision-enabled chat models, it's essential to tailor the prompts to your specific needs. Here are some guidelines to enhance the accuracy and efficiency of your prompts.
Copy file name to clipboardExpand all lines: articles/ai-services/openai/concepts/gpt-with-vision.md
+12-12Lines changed: 12 additions & 12 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -6,13 +6,13 @@ author: PatrickFarley
6
6
ms.author: pafarley
7
7
ms.service: azure-ai-openai
8
8
ms.topic: conceptual
9
-
ms.date: 02/20/2025
9
+
ms.date: 05/31/2025
10
10
manager: nitinme
11
11
---
12
12
13
13
# Vision-enabled chat model concepts
14
14
15
-
Vision-enabled chat models are large multimodal models (LMM) developed by OpenAI that can analyze images and provide textual responses to questions about them. They incorporate both natural language processing and visual understanding. The current vision-enabled models are GPT-4 Turbo with Vision, GPT-4o, and GPT-4o-mini. This guide provides details on their capabilities and limitations.
15
+
Vision-enabled chat models are large multimodal models (LMM) developed by OpenAI that can analyze images and provide textual responses to questions about them. They incorporate both natural language processing and visual understanding. This guide provides details on their capabilities and limitations. To see which models support image input, see the [Models page](./models.md).
16
16
17
17
To try out vision-enabled chat models, see the [quickstart](/azure/ai-services/openai/gpt-v-quickstart).
18
18
@@ -21,6 +21,16 @@ To try out vision-enabled chat models, see the [quickstart](/azure/ai-services/o
21
21
The vision-enabled models answer general questions about what's present in the images you upload.
22
22
23
23
24
+
## Input limitations
25
+
26
+
This section describes the limitations of vision-enabled chat models.
27
+
28
+
### Image support
29
+
30
+
-**Maximum input image size**: The maximum size for input images is restricted to 20 MB.
31
+
-**Low resolution accuracy**: When images are analyzed using the "low resolution" setting, it allows for faster responses and uses fewer input tokens for certain use cases. However, this could impact the accuracy of object and text recognition within the image.
32
+
-**Image chat restriction**: When you upload images in [Azure AI Foundry portal](https://ai.azure.com/) or the API, there is a limit of 10 images per chat call.
33
+
24
34
## Special pricing information
25
35
26
36
> [!IMPORTANT]
@@ -70,16 +80,6 @@ For a typical use case, take a 3-minute video with a 100-token prompt input. The
70
80
Additionally, there's a one-time indexing cost of $0.15 to generate the Video Retrieval index for this 3-minute video. This index can be reused across any number of Video Retrieval and GPT-4 Turbo with Vision API calls.
71
81
-->
72
82
73
-
## Input limitations
74
-
75
-
This section describes the limitations of vision-enabled chat models.
76
-
77
-
### Image support
78
-
79
-
-**Maximum input image size**: The maximum size for input images is restricted to 20 MB.
80
-
-**Low resolution accuracy**: When images are analyzed using the "low resolution" setting, it allows for faster responses and uses fewer input tokens for certain use cases. However, this could impact the accuracy of object and text recognition within the image.
81
-
-**Image chat restriction**: When you upload images in [Azure AI Foundry portal](https://ai.azure.com/) or the API, there is a limit of 10 images per chat call.
Get started using GPT-4 Turbo with images with the Azure OpenAI in Azure AI Foundry Models.
18
+
Get started using images in your chats with Azure OpenAI in Azure AI Foundry Models.
19
19
20
-
> [!NOTE]
21
-
> **Model choice**
22
-
>
23
-
> The latest vision-capable models are `gpt-4o` and `gpt-4o mini`. These models are in public preview. The latest available GA model is `gpt-4` version `turbo-2024-04-09`.
24
20
25
21
> [!IMPORTANT]
26
22
> Extra usage fees might apply when using chat completion models with vision functionality.
@@ -62,7 +58,7 @@ Get started using GPT-4 Turbo with images with the Azure OpenAI in Azure AI Foun
62
58
63
59
::: zone-end
64
60
65
-
## Next steps
61
+
## Related content
66
62
67
63
*[Get started with multimodal vision chat apps using Azure OpenAI AI App template](/azure/developer/ai/get-started-app-chat-vision?tabs=github-codespaces)
68
64
* Learn more about these APIs in the [Vision-enabled models how-to guide](./gpt-v-quickstart.md)
Copy file name to clipboardExpand all lines: articles/ai-services/openai/how-to/gpt-with-vision.md
+3-3Lines changed: 3 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,14 +7,14 @@ ms.author: pafarley #delegenz
7
7
#customer intent: As a developer, I want to learn how to use vision-enabled chat models so that I can integrate image processing capabilities into my applications.
8
8
ms.service: azure-ai-openai
9
9
ms.topic: how-to
10
-
ms.date: 02/20/2025
10
+
ms.date: 05/31/2025
11
11
manager: nitinme
12
12
---
13
13
14
14
# Use vision-enabled chat models
15
15
16
16
17
-
Vision-enabled chat models are large multimodal models (LMM) developed by OpenAI that can analyze images and provide textual responses to questions about them. They incorporate both natural language processing and visual understanding. The current vision-enabled models are [o1](./reasoning.md), GPT-4o, GPT-4o-mini, and GPT-4 Turbo with Vision.
17
+
Vision-enabled chat models are large multimodal models (LMM) developed by OpenAI that can analyze images and provide textual responses to questions about them. They incorporate both natural language processing and visual understanding. The current vision-enabled models are the [o-series reasoning models](./reasoning.md), GPT-4.1 series models, GPT-4.5, GPT-4o series, and GPT-4 Turbo with Vision.
18
18
19
19
The vision-enabled models can answer general questions about what's present in the images you upload.
20
20
@@ -383,7 +383,7 @@ Every response includes a `"finish_reason"` field. It has the following possible
Copy file name to clipboardExpand all lines: articles/ai-services/openai/how-to/risks-safety-monitor.md
+3-3Lines changed: 3 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -6,7 +6,7 @@ author: PatrickFarley
6
6
ms.author: pafarley
7
7
ms.service: azure-ai-openai
8
8
ms.topic: how-to
9
-
ms.date: 02/20/2025
9
+
ms.date: 05/31/2025
10
10
manager: nitinme
11
11
---
12
12
@@ -43,14 +43,14 @@ Adjust your content filter configuration to further align with business needs an
43
43
44
44
## Potentially abusive user detection
45
45
46
-
The **Potentially abusive user detection** pane leverages user-level abuse reporting to show information about users whose behavior has resulted in blocked content. The goal is to help you get a view of the sources of harmful content so you can take responsive actions to ensure the model is being used in a responsible way.
46
+
The **Potentially abusive user detection** pane shows information about users whose behavior has resulted in blocked content. The goal is to help you get a view of the sources of harmful content so you can take responsive actions to ensure the model is being used in a responsible way.
47
47
48
48
49
49
To use Potentially abusive user detection, you need:
50
50
- A content filter configuration applied to your deployment.
51
51
- You must be sending user ID information in your Chat Completion requests (see the _user_ parameter of the [Completions API](/azure/ai-services/openai/reference#completions), for example).
52
52
> [!CAUTION]
53
-
> Use GUID strings to identify individual users. Do not include sensitive personal information in the _user_ field.
53
+
> Use GUID strings to identify individual users. Don't include sensitive personal information in the _user_ field.
54
54
- An Azure Data Explorer database set up to store the user analysis results (instructions below).
Copy file name to clipboardExpand all lines: articles/ai-services/openai/how-to/use-blocklists.md
+5-5Lines changed: 5 additions & 5 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -5,7 +5,7 @@ description: Learn how to use blocklists with Azure OpenAI
5
5
manager: nitinme
6
6
ms.service: azure-ai-openai
7
7
ms.topic: how-to
8
-
ms.date: 02/20/2025
8
+
ms.date: 05/31/2025
9
9
author: PatrickFarley
10
10
ms.author: pafarley
11
11
---
@@ -17,7 +17,7 @@ The [configurable content filters](/azure/ai-services/openai/how-to/content-filt
17
17
## Prerequisites
18
18
19
19
- An Azure subscription. <ahref="https://azure.microsoft.com/free/ai-services"target="_blank">Create one for free</a>.
20
-
- Once you have your Azure subscription, create an Azure OpenAI resource in the Azure portal to get your token, key and endpoint. Enter a unique name for your resource, select the subscription you entered on the application form, select a resource group, supported region, and supported pricing tier. Then select **Create**.
20
+
- Once you have your Azure subscription, create an Azure OpenAI resource in the Azure portal to get your token, key, and endpoint. Enter a unique name for your resource, select the subscription you entered on the application form, select a resource group, supported region, and supported pricing tier. Then select **Create**.
21
21
- The resource takes a few minutes to deploy. After it finishes, select **go to resource**. In the left pane, under **Resource Management**, select **Subscription Key and Endpoint**. The endpoint and either of the keys are used to call APIs.
@@ -30,7 +30,7 @@ You can create blocklists with the Azure OpenAI API. The following steps help yo
30
30
31
31
### Get your token
32
32
33
-
First, you need to get a token for accessing the APIs for creating, editing and deleting blocklists. You can get this token using the following Azure CLI command:
33
+
First, you need to get a token for accessing the APIs for creating, editing, and deleting blocklists. You can get this token using the following Azure CLI command:
34
34
35
35
```bash
36
36
az account get-access-token
@@ -102,7 +102,7 @@ Copy the cURL command below to a text editor and make the following changes:
102
102
1. Replace {raiBlocklistName} (in the URL) with a custom name for your list. Allowed characters: `0-9, A-Z, a-z, - . _ ~`.
103
103
1. Replace {raiBlocklistItemName} with a custom name for your list item.
104
104
1. Replace {token} with the token you got from the "Get your token" step above.
105
-
1. Replace the value of the `"blocking pattern"` field with the item you'd like to add to your blocklist. The maximum length of a blockItem is 1000 characters. Also specify whether the pattern is regex or exact match.
105
+
1. Replace the value of the `"blocking pattern"` field with the item you'd like to add to your blocklist. The maximum length of a blockItem is 1,000 characters. Also specify whether the pattern is regex or exact match.
106
106
107
107
```bash
108
108
curl --location --request PUT 'https://management.azure.com/subscriptions/{subscriptionId}/resourceGroups/{resourceGroupName}/providers/Microsoft.CognitiveServices/accounts/{accountName}/raiBlocklists/{raiBlocklistName}/raiBlocklistItems/{raiBlocklistItemName}?api-version=2024-04-01-preview'\
@@ -117,7 +117,7 @@ curl --location --request PUT 'https://management.azure.com/subscriptions/{subsc
117
117
```
118
118
119
119
> [!NOTE]
120
-
> It can take around 5 minutes for a new term to be added to the blocklist. Please test after 5 minutes.
120
+
> It can take around 5 minutes for a new term to be added to the blocklist. Test the blocklist after 5 minutes.
0 commit comments