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
@@ -18,7 +18,7 @@ Azure OpenAI Service is powered by a diverse set of models with different capabi
18
18
19
19
| Models | Description |
20
20
|--|--|
21
-
|[o1 & o1-mini](#o1-and-o1-mini-models)|[Reasoning models](../how-to/reasoning.md) with advanced problem-solving and increased focus and capability. |
21
+
|[o-series models](#o-series-models)|[Reasoning models](../how-to/reasoning.md) with advanced problem-solving and increased focus and capability. |
22
22
|[GPT-4o & GPT-4o mini & GPT-4 Turbo](#gpt-4o-and-gpt-4-turbo)| The latest most capable Azure OpenAI models with multimodal versions, which can accept both text and images as input. |
23
23
|[GPT-4o audio](#gpt-4o-audio)| GPT-4o audio models that support either low-latency, "speech in, speech out" conversational interactions or audio generation. |
24
24
|[GPT-4](#gpt-4)| A set of models that improve on GPT-3.5 and can understand and generate natural language and code. |
@@ -28,30 +28,32 @@ Azure OpenAI Service is powered by a diverse set of models with different capabi
28
28
|[Whisper](#whisper-models)| A series of models in preview that can transcribe and translate speech to text. |
29
29
|[Text to speech](#text-to-speech-models-preview) (Preview) | A series of models in preview that can synthesize text to speech. |
30
30
31
-
## o1 and o1-mini models
31
+
## o-series models
32
32
33
-
The Azure OpenAI `o1` and `o1-mini` models are specifically designed to tackle reasoning and problem-solving tasks with increased focus and capability. These models spend more time processing and understanding the user's request, making them exceptionally strong in areas like science, coding, and math compared to previous iterations.
33
+
The Azure OpenAI o<sup>*</sup> series models are specifically designed to tackle reasoning and problem-solving tasks with increased focus and capability. These models spend more time processing and understanding the user's request, making them exceptionally strong in areas like science, coding, and math compared to previous iterations.
34
34
35
35
| Model ID | Description | Max Request (tokens) | Training Data (up to) |
|`o1` (2024-12-17) | The most capable model in the o1 series, offering [enhanced reasoning abilities](../how-to/reasoning.md). <br> - Structured outputs<br> - Text, image processing <br> - Functions/Tools <br> <br> **Request access: [limited access model application](https://aka.ms/OAI/o1access)**| Input: 200,000 <br> Output: 100,000 | Oct 2023 |
38
39
|`o1-preview` (2024-09-12) | Older preview version | Input: 128,000 <br> Output: 32,768 | Oct 2023 |
39
40
|`o1-mini` (2024-09-12) | A faster and more cost-efficient option in the o1 series, ideal for coding tasks requiring speed and lower resource consumption.| Input: 128,000 <br> Output: 65,536 | Oct 2023 |
40
41
41
42
### Availability
42
43
43
-
The `o1` and `o1-mini`models are now available for API access and model deployment. **For access to `o1` registration is required, and access will be granted based on Microsoft's eligibility criteria**. Customers who previously applied and received access to `o1-preview`, don't need to reapply as they are automatically on the wait-list for the latest model.
44
+
**For access to `o3-mini` and `o1` registration is required, and access will be granted based on Microsoft's eligibility criteria**. Customers who previously applied and received access to `o1-preview`, or `o1`don't need to reapply as they are automatically on the wait-list for the latest models in the o-series.
44
45
45
46
Request access: [limited access model application](https://aka.ms/OAI/o1access)
46
47
47
-
Once access has been granted, you will need to create a deployment for each model. If you have an existing `o1-preview` deployment in place upgrade is currently not supported, you will need to create a new deployment.
48
+
Once access has been granted, you will need to create a deployment for each model.
48
49
49
-
To learn more about the advanced `o1` series models see, [getting started with o1 series reasoning models](../how-to/reasoning.md).
50
+
To learn more about the advanced `o-series`models see, [getting started with reasoning models](../how-to/reasoning.md).
50
51
51
52
### Region availability
52
53
53
54
| Model | Region |
54
55
|---|---|
56
+
|`o3-mini`| East US2 (Global Standard) <br> Sweden Central (Global Standard) |
55
57
|`o1`| East US2 (Global Standard) <br> Sweden Central (Global Standard) |
56
58
|`o1-preview`| See the [models table](#global-standard-model-availability). |
57
59
|`o1-mini`| See the [models table](#global-provisioned-managed-model-availability). |
Copy file name to clipboardExpand all lines: articles/ai-services/openai/how-to/function-calling.md
+4-3Lines changed: 4 additions & 3 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -7,7 +7,7 @@ ms.author: mbullwin #delegenz
7
7
ms.service: azure-ai-openai
8
8
ms.custom: devx-track-python
9
9
ms.topic: how-to
10
-
ms.date: 01/17/2025
10
+
ms.date: 01/30/2025
11
11
manager: nitinme
12
12
---
13
13
@@ -45,14 +45,15 @@ Support for parallel function was first added in API version [`2023-12-01-previe
45
45
### Basic function calling with tools
46
46
47
47
* All the models that support parallel function calling
48
+
*`o3-mini` (`2025-01-31`)
48
49
*`o1` (`2024-12-17`)
49
50
*`gpt-4` (`0613`)
50
51
*`gpt-4-32k` (`0613`)
51
52
*`gpt-35-turbo-16k` (`0613`)
52
53
*`gpt-35-turbo` (`0613`)
53
54
54
-
> [!IMPORTANT]
55
-
> There is a known issue with the `o1` model and the `tool_choice` parameter. Currently function calls that include the optional `tool_choice` parameter will fail. This page will be updated once the issue is resolved. For more information on what parameters are supported with the o1-series models see, the [reasoning models guide](./reasoning.md).
55
+
> [!NOTE]
56
+
> The `tool_choice` parameter is now supported with `o3-mini` and `o1`. For more information on what parameters are supported with the o-series models see, the [reasoning models guide](./reasoning.md).
Copy file name to clipboardExpand all lines: articles/ai-services/openai/how-to/reasoning.md
+23-25Lines changed: 23 additions & 25 deletions
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -5,17 +5,17 @@ description: Learn how to use Azure OpenAI's advanced o1 series reasoning models
5
5
manager: nitinme
6
6
ms.service: azure-ai-openai
7
7
ms.topic: include
8
-
ms.date: 01/16/2025
8
+
ms.date: 01/30/2025
9
9
author: mrbullwinkle
10
10
ms.author: mbullwin
11
11
---
12
12
13
13
14
14
# Azure OpenAI reasoning models
15
15
16
-
Azure OpenAI `o1` and `o1-mini` models are designed to tackle reasoning and problem-solving tasks with increased focus and capability. These models spend more time processing and understanding the user's request, making them exceptionally strong in areas like science, coding, and math compared to previous iterations.
16
+
Azure OpenAI `o-series` models are designed to tackle reasoning and problem-solving tasks with increased focus and capability. These models spend more time processing and understanding the user's request, making them exceptionally strong in areas like science, coding, and math compared to previous iterations.
17
17
18
-
**Key capabilities of the o1 series:**
18
+
**Key capabilities of the o-series models:**
19
19
20
20
- Complex Code Generation: Capable of generating algorithms and handling advanced coding tasks to support developers.
21
21
- Advanced Problem Solving: Ideal for comprehensive brainstorming sessions and addressing multifaceted challenges.
@@ -24,45 +24,43 @@ Azure OpenAI `o1` and `o1-mini` models are designed to tackle reasoning and prob
24
24
25
25
## Availability
26
26
27
-
The**o1 series** models are now available for API access and model deployment. **For access to o1, and o1-preview registration is required, and access will be granted based on Microsoft's eligibility criteria**. Customers who previously applied and received access to `o1-preview`, don't need to reapply as they are automatically on the wait-list for the latest model.
27
+
**For access to `o3-mini`, `o1`, and `o1-preview`, registration is required, and access will be granted based on Microsoft's eligibility criteria**.
28
28
29
-
Request access: [limited access model application](https://aka.ms/OAI/o1access)
29
+
Customers who previously applied and received access to `o1` or `o1-preview`, don't need to reapply as they are automatically on the wait-list for the latest model.
30
30
31
-
Once access has been granted, you'll need to create a deployment for each model. If you have an existing `o1-preview` deployment, in-place upgrade is currently not supported, you'll need to create a new deployment.
31
+
Request access: [limited access model application](https://aka.ms/OAI/o1access)
32
32
33
33
### Region availability
34
34
35
35
| Model | Region | Limited access |
36
36
|---|---|---|
37
+
|`o3-mini`| East US2 (Global Standard) <br> Sweden Central (Global Standard) |[Limited access model application](https://aka.ms/OAI/o1access)|
37
38
|`o1`| East US2 (Global Standard) <br> Sweden Central (Global Standard) |[Limited access model application](https://aka.ms/OAI/o1access)|
38
39
|`o1-preview`| See [models page](../concepts/models.md#global-standard-model-availability). |[Limited access model application](https://aka.ms/OAI/o1access)|
39
40
|`o1-mini`| See [models page](../concepts/models.md#global-standard-model-availability). | No access request needed |
**o1 series** models will only work with the `max_completion_tokens` parameter.
56
-
57
-
> [!IMPORTANT]
58
-
> There is a known issue with the `o1` model and the `tool_choice` parameter. Currently function calls that include the optional `tool_choice` parameter will fail. This page will be updated once the issue is resolved.
<sup>*</sup> Reasoning models will only work with the `max_completion_tokens` parameter. <br><br>
58
+
<sup>**</sup>The latest o<sup>*</sup> series model support system messages to make migration easier. When you use a system message with `o3-mini` and `o1` it will be treated as a developer message. You should not use both a developer message and a system message in the same API request.
59
59
60
60
### Not Supported
61
61
62
-
The following are currently unsupported with o1-series models:
62
+
The following are currently unsupported with reasoning models:
Copy file name to clipboardExpand all lines: articles/ai-services/openai/overview.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
@@ -7,20 +7,20 @@ author: mrbullwinkle
7
7
ms.author: mbullwin
8
8
ms.service: azure-ai-openai
9
9
ms.topic: overview
10
-
ms.date: 01/23/2025
10
+
ms.date: 01/30/2025
11
11
ms.custom: build-2023, build-2023-dataai
12
12
recommendations: false
13
13
---
14
14
15
15
# What is Azure OpenAI Service?
16
16
17
-
Azure OpenAI Service provides REST API access to OpenAI's powerful language models including o1, o1-mini, GPT-4o, GPT-4o mini, GPT-4 Turbo with Vision, GPT-4, GPT-3.5-Turbo, and Embeddings model series. These models can be easily adapted to your specific task including but not limited to content generation, summarization, image understanding, semantic search, and natural language to code translation. Users can access the service through REST APIs, Python SDK, or in the [Azure AI Foundry](https://ai.azure.com).
17
+
Azure OpenAI Service provides REST API access to OpenAI's powerful language models including o3-mini, o1, o1-mini, GPT-4o, GPT-4o mini, GPT-4 Turbo with Vision, GPT-4, GPT-3.5-Turbo, and Embeddings model series. These models can be easily adapted to your specific task including but not limited to content generation, summarization, image understanding, semantic search, and natural language to code translation. Users can access the service through REST APIs, Python SDK, or in the [Azure AI Foundry](https://ai.azure.com).
18
18
19
19
### Features overview
20
20
21
21
| Feature | Azure OpenAI |
22
22
| --- | --- |
23
-
| Models available |[**o1**](./how-to/reasoning.md) - (Limited Access - [Request Access](https://aka.ms/OAI/o1access))<br>[**o1-mini**](./how-to/reasoning.md)<br>**GPT-4o & GPT-4o mini**<br> **GPT-4 series (including GPT-4 Turbo with Vision)** <br>**GPT-3.5-Turbo series**<br> Embeddings series <br> Learn more in our [Models](./concepts/models.md) page.|
23
+
| Models available |[**o3-mini & o1**](./how-to/reasoning.md) - (Limited Access - [**Request Access**](https://aka.ms/OAI/o1access))<br>[**o1-mini**](./how-to/reasoning.md)<br>**GPT-4o & GPT-4o mini**<br> **GPT-4 series (including GPT-4 Turbo with Vision)** <br>**GPT-3.5-Turbo series**<br> Embeddings series <br> Learn more in our [Models](./concepts/models.md) page.|
| Price |[Available here](https://azure.microsoft.com/pricing/details/cognitive-services/openai-service/) <br> For details on vision-enabled chat models, see the [special pricing information](../openai/concepts/gpt-with-vision.md#special-pricing-information).|
26
26
| Virtual network support & private link support | Yes. |
> The ratio of RPM/TPM for quota with o1-series models works differently than older chat completions models:
69
69
>
70
70
> -**Older chat models:** 1 unit of capacity = 6 RPM and 1,000 TPM.
71
71
> -**o1 & o1-preview:** 1 unit of capacity = 1 RPM and 6,000 TPM.
72
+
> -**o3-mini:** 1 unit of capacity = 1 RPM per 10,000 TPM.
72
73
> -**o1-mini:** 1 unit of capacity = 1 RPM per 10,000 TPM.
73
74
>
74
75
> This is particularly important for programmatic model deployment as this change in RPM/TPM ratio can result in accidental under allocation of quota if one is still assuming the 1:1000 ratio followed by older chat completion models.
75
76
>
76
77
> There is a known issue with the [quota/usages API](/rest/api/aiservices/accountmanagement/usages/list?view=rest-aiservices-accountmanagement-2024-06-01-preview&tabs=HTTP&preserve-view=true) where it assumes the old ratio applies to the new o1-series models. The API returns the correct base capacity number, but doesn't apply the correct ratio for the accurate calculation of TPM.
77
78
78
-
### o1 & o1-mini global standard
79
+
### `o-series` global standard
79
80
80
81
| Model|Tier| Quota Limit in tokens per minute (TPM) | Requests per minute |
81
82
|---|---|:---:|:---:|
83
+
|`o3-mini`| Enterprise agreement | 50 M | 5 K |
82
84
|`o1` & `o1-preview`| Enterprise agreement | 30 M | 5 K |
Copy file name to clipboardExpand all lines: articles/ai-services/openai/whats-new.md
+5-1Lines changed: 5 additions & 1 deletion
Display the source diff
Display the rich diff
Original file line number
Diff line number
Diff line change
@@ -11,7 +11,7 @@ ms.custom:
11
11
- references_regions
12
12
- ignite-2024
13
13
ms.topic: whats-new
14
-
ms.date: 1/21/2025
14
+
ms.date: 1/30/2025
15
15
recommendations: false
16
16
---
17
17
@@ -21,6 +21,10 @@ This article provides a summary of the latest releases and major documentation u
21
21
22
22
## January 2025
23
23
24
+
### o3-mini released
25
+
26
+
`o3-mini` (2025-01-31) is the latest reasoning model, offering enhanced reasoning abilities. For more information, see our [reasoning model guide](./how-to/reasoning.md).
27
+
24
28
### GPT-4o audio completions
25
29
26
30
The `gpt-4o-audio-preview` model is now available for global deployments in [East US 2 and Sweden Central regions](./concepts/models.md#global-standard-model-availability). Use the `gpt-4o-audio-preview` model for audio generation.
0 commit comments