Merge pull request #4475 from MicrosoftDocs/main

v-alje · web-flow · commit cfed4022a868 · 2025-05-01T15:33:04.000-07:00
05/01/2025 PM Publishing
diff --git a/articles/ai-services/content-understanding/concepts/analyzer-templates.md b/articles/ai-services/content-understanding/concepts/analyzer-templates.md
@@ -7,7 +7,7 @@ ms.author: kabrow
 manager: nitinme
 ms.service: azure-ai-content-understanding
 ms.topic: overview
-ms.date: 02/19/2025
+ms.date: 05/01/2025
 ms.custom: ignite-2024-understanding-release
 ---
 
@@ -26,7 +26,7 @@ The following analyzer templates are available for use in the [Azure AI Foundry
 |Template| Description|
 | ----|----|----|
 |Document analysis |Analyze documents to extract text, layout, structured fields, and more.|
-|Text analysis |Analyze texts and extract structured fields.|
+|Invoice analysis |Analyze invoice as prebuilt template and extract structured fields and tables.|
 
    :::image type="content" source="../media/analyzer-template/scenario-template-document.png" alt-text="Screenshot of document analyzer template.":::
 
diff --git a/articles/ai-services/content-understanding/document/overview.md b/articles/ai-services/content-understanding/document/overview.md
@@ -7,7 +7,7 @@ ms.author: lajanuar
 manager: nitinme
 ms.service: azure-ai-content-understanding
 ms.topic: overview
-ms.date: 02/19/2025
+ms.date: 05/01/2025
 ms.custom: ignite-2024-understanding-release
 ---
 
@@ -23,6 +23,12 @@ Content Understanding is a cloud-based [Azure AI Service](../../what-are-ai-serv
 
 Content Understanding enables organization to streamline data collection and processing, enhance operational efficiency, optimize data-driven decision making, and empower innovation. With customizable analyzers, Content Understanding allows for easy extraction of content or fields from documents and forms, tailored to specific business needs.
 
+## April updates
+
+* **Invoice prebuilt template**: Extract predefined schemas from various invoice formats. The out-of-the-box schema can be customized by adding or removing fields to suit your specific needs.
+
+* **Generative and classify methods**: Added support for both generative and classification-based methods, enabling you to create generative fields such as summaries or categorize document details into multiple classes using the classify method.
+
 ## Business use cases
 
 Document analyzers can process complex documents in various formats and templates:
diff --git a/articles/ai-services/content-understanding/service-limits.md b/articles/ai-services/content-understanding/service-limits.md
@@ -69,11 +69,11 @@ The following limits apply as of version 2024-12-01-preview.
 
 | Property | Document | Image | Text | Audio | Video |
 | --- | --- | --- | --- | --- | --- |
-| Max fields | 50 | 10 | 10 | 10 | 10 |
+| Max fields | 50 | 50 | 50 | 50 | 50 |
 | Max classify field categories | 300 | 300 | 300 | 300 | 300 |
 | Supported generation methods | extract | generate<br>classify | generate<br>classify | generate<br>classify | generate<br>classify |
 
-* The *Max fields* limit includes all named fields. For example, a list of strings counts as one field, while a group with string and number subfields counts as three fields. To extract beyond default limits, contact us at cu_contact@microsoft.com.
+* The *Max fields* limit includes all named fields. For example, a list of strings counts as one field, while a group with string and number subfields counts as three fields. To extend the limit for documents fields up to 100, contact us at `cu_contact@microsoft.com`.
 * The *Max classify field categories* limit is the total number of categories across all fields using the `classify` generation method.
 * The generation method currently applies only to basic fields.
 
@@ -83,7 +83,7 @@ The following limits apply as of version 2024-12-01-preview.
 | --- | --- | --- | --- | --- | --- |
 | Basic | No *boolean* | No *date*, *time* | *string* | *string* | No *date*, *time* |
 | List | N/A | No *date*, *time* | *string* | *string* | No *date*, *time* |
-| Group | N/A | No *date*, *time* | N/A | N/A | No *date*, *time* |
+| Group | N/A | No *date*, *time* |*string* | *string* | No *date*, *time* |
 | Table | No *boolean* | No *date*, *time* | *string* | *string* | No *date*, *time* |
 | Fixed table | No *boolean* | N/A | N/A | N/A | N/A |
 
diff --git a/articles/ai-services/content-understanding/whats-new.md b/articles/ai-services/content-understanding/whats-new.md
@@ -41,4 +41,5 @@ The Content Understanding **2024-12-01-preview** REST API is now available. This
   * Add download code samples for quick setup added.
 
 ## November 2024
-Welcome! The Azure AI Content Understanding API version `2024-12-01-preview` is now in public preview. This version allows you to generate a structured representation of content tailored to specific tasks from various modalities or formats. Content Understanding uses a defined schema to extract content suitable for processing by large language models and subsequent applications.
+
+Welcome! The Azure AI Content Understanding API version `2024-12-01-preview` is now in public preview. This version allows you to generate a structured representation of content tailored to specific tasks from various modalities or formats. Content Understanding uses a defined schema to extract content suitable for processing by large language models and subsequent applications.
diff --git a/articles/ai-services/openai/concepts/models.md b/articles/ai-services/openai/concepts/models.md
@@ -43,7 +43,7 @@ Azure OpenAI Service is powered by a diverse set of models with different capabi
 
 |  Model ID  | Description | Context Window | Max Output Tokens | Training Data (up to)  |
 |  --- |  :--- |:--- |:---|:---: |
-| `gpt-4.1` (2025-04-14)   | - Text & image input <br> - Text output <br> - Chat completions API <br>- Responses API <br> - Streaming <br> - Function calling <br> Structured outputs (chat completions)   | 1,047,576 | 32,768 | May 31, 2024 |
+| `gpt-4.1` (2025-04-14)   | - Text & image input <br> - Text output <br> - Chat completions API <br>- Responses API <br> - Streaming <br> - Function calling <br> Structured outputs (chat completions)   | - 1,047,576 <br> - 128,000 (provisioned managed deployments) | 32,768 | May 31, 2024 |
 | `gpt-4.1-nano` (2025-04-14) <br><br> **Fastest 4.1 model** | - Text & image input <br> - Text output <br> - Chat completions API <br>- Responses API <br> - Streaming <br> - Function calling <br> Structured outputs (chat completions)   | 1,047,576  | 32,768 | May 31, 2024 |
 | `gpt-4.1-mini` (2025-04-14) | - Text & image input <br> - Text output <br> - Chat completions API <br>- Responses API <br> - Streaming <br> - Function calling <br> Structured outputs (chat completions)   | 1,047,576  | 32,768 | May 31, 2024 |
 
diff --git a/articles/ai-services/openai/concepts/provisioned-throughput.md b/articles/ai-services/openai/concepts/provisioned-throughput.md
@@ -3,7 +3,7 @@ title: Azure OpenAI Service provisioned throughput
 description: Learn about provisioned throughput and Azure OpenAI.
 ms.service: azure-ai-openai
 ms.topic: conceptual
-ms.date: 03/31/2025
+ms.date: 04/30/2025
 manager: nitinme
 author: aahill #ChrisHMSFT
 ms.author: aahi #chrhoder
diff --git a/articles/ai-services/openai/how-to/provisioned-throughput-onboarding.md b/articles/ai-services/openai/how-to/provisioned-throughput-onboarding.md
@@ -72,23 +72,24 @@ Customers that require long-term usage of provisioned, data zoned provisioned, a
 > Charges for deployments on a deleted resource will continue until the resource is purged. To prevent this, delete a resource’s deployment before deleting the resource. For more information, see [Recover or purge deleted Azure AI services resources](../../recover-purge-resources.md). 
 
 ## How much throughput per PTU you get for each model
-The amount of throughput (measured in tokens per minute or TPM) a deployment gets per PTU is a function of the input and output tokens in a given minute. 
 
-Generating output tokens requires more processing than input tokens. For the models specified in the table below, 1 output token counts as 3 input tokens towards your TPM-per-PTU limit. The service dynamically balances the input & output costs, so users do not have to set specific input and output limits. This approach means your deployment is resilient to fluctuations in the workload.
 
-To help with simplifying the sizing effort, the following table outlines the TPM-per-PTU for the specified models. To understand the impact of output tokens on the TPM-per-PTU limit, use the 3 input token to 1 output token ratio. 
 
-For a detailed understanding of how different ratios of input and output tokens impact the throughput your workload needs, see the [Azure OpenAI capacity calculator](https://ai.azure.com/resource/calculator). The table also shows Service Level Agreement (SLA) Latency Target Values per model. For more information about the SLA for Azure OpenAI Service, see the [Service Level Agreements (SLA) for Online Services page](https://www.microsoft.com/licensing/docs/view/Service-Level-Agreements-SLA-for-Online-Services?lang=1)
 
 
-|Topic| **gpt-4o**   | **gpt-4o-mini**  | **o1**|
-| --- | --- | --- | --- |
-|Global & data zone provisioned minimum deployment|15|15|15|
-|Global & data zone provisioned scale increment|5|5|5|
-|Regional provisioned minimum deployment|50|25|50|
-|Regional provisioned scale increment|50|25|50|
-|Input TPM per PTU |2,500|37,000|230|
-|Latency Target Value |25 Tokens Per Second|33 Tokens Per Second|25 Tokens Per Second|
+The amount of throughput (measured in tokens per minute or TPM) a deployment gets per PTU is a function of the input and output tokens in a given minute. Generating output tokens requires more processing than input tokens.  Starting with GPT 4.1 models and later, the system matches the global standard price ratio between input and output tokens. Cached tokens are deducted 100% from the utilization.
+
+For example, for `gpt-4.1:2025-04-14`, 1 output token counts as 4 input tokens towards your utilization limit which matches the [pricing](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/). Older models use a different ratio and for a deeper understanding on how different ratios of input and output tokens impact the throughput your workload needs, see the [Azure OpenAI capacity calculator](https://ai.azure.com/resource/calculator).
+
+
+|Topic| **gpt-4o**   | **gpt-4o-mini**  | **o1**| gpt-4.1 |
+| --- | --- | --- | --- | --- |
+|Global & data zone provisioned minimum deployment|15|15|15|15 |
+|Global & data zone provisioned scale increment|5|5|5| 5 | 
+|Regional provisioned minimum deployment|50|25|50| 50 |
+|Regional provisioned scale increment|50|25|50| 50 |
+|Input TPM per PTU |2,500|37,000|230| 3000 | 
+|Latency Target Value |25 Tokens Per Second|33 Tokens Per Second|25 Tokens Per Second| 44 Tokens Per Second |
 
 For a full list, see the [Azure OpenAI Service in Azure AI Foundry portal calculator](https://ai.azure.com/resource/calculator).
 
diff --git a/articles/ai-services/openai/includes/model-matrix/provisioned-global.md b/articles/ai-services/openai/includes/model-matrix/provisioned-global.md
