GPAI Ledger › Phi 4 Reasoning Vision 15B (Microsoft) › Capture 13 Sep 2026

Phi 4 Reasoning Vision 15B — capture 20260913T121419Z

Verify: sha256sum 7bc0e89f1dcfeb9817d31a67b255443a0457082c3ae8fbaf226d6c31d31cd0a9.md must equal the hash above (the filename IS the expected hash); ots verify 7bc0e89f1dcfeb9817d31a67b255443a0457082c3ae8fbaf226d6c31d31cd0a9.md.20260913T121419Z.ots -f 7bc0e89f1dcfeb9817d31a67b255443a0457082c3ae8fbaf226d6c31d31cd0a9.md (opentimestamps.org) proves the bytes existed no later than the attestation time — an upper bound on the capture time; the fetch time above is the archive's own record (a freshly captured proof reports 'pending' here: the calendars anchor within hours, but this archive only upgrades the stored proof to its anchor on a later run, so expect a day or two). ots verify needs a local Bitcoin Core node (a pruned one is fine); without one, ots info on the proof prints the attesting block height and merkle path to check on any block explorer.

Extracted text

Machine-extracted text (layout may be lost; the authoritative content is the stored file above).

# Phi-4-Reasoning-Vision-15B

**Version of the Summary:**

*1.0*

**Last update:** 31 Aug 2026

## 1. General information

### 1.1 Model developer identification

**1.1.1 Model developer name and contact details:** Microsoft Ireland
Operations Limited (MIOL) 70 Sir John Rogerson’s Quay, Dublin 2, D02
R296, Ireland

**1.1.2 Authorized representative name and contact details:**
<MSFTAIActRequest@microsoft.com>

### 1.2: Model identification

**1.2.1 Versioned model name(s):** Phi-4-Reasoning-Vision-15B

**1.2.2 Model dependencies:** Phi-4-Reasoning

**1.2.3 Model release date:** 4-March-2026

**1.2.4 Date of placement of the model on the Union market:**
4-March-2026

### 1.3 Modalities, overall training data size and other characteristics

**1.3.1 Size of dataset and characteristics**

<table>
<colgroup>
<col style="width: 28%" />
<col style="width: 32%" />
<col style="width: 39%" />
</colgroup>
<thead>
<tr class="header">
<th><blockquote>
<p><strong>Modality</strong></p>
</blockquote></th>
<th><strong>Training data size</strong></th>
<th><strong>Types of content</strong></th>
</tr>
</thead>
<tbody>
<tr class="odd">
<td><blockquote>
<p>☒Text</p>
</blockquote></td>
<td><p>☐Less than 1 billion tokens</p>
<p>☒1billion to10 trillions tokens</p>
<p>☐More than 10 trillions tokens</p>
<blockquote>
<p>Alternatively, specify the approximate size in
a different measurement unit:</p>
</blockquote></td>
<td><blockquote>
<p><em>Image captions of varying detail, question-answer pairs,
step-by-step math solutions with LaTeX, OCR transcriptions, safety
refusals, multi-turn tutoring dialogues, document QA, UI action
instructions, HTML/code, diagram and chart interpretations, and
chain-of-thought reasoning traces.</em></p>
</blockquote></td>
</tr>
<tr class="even">
<td><blockquote>
<p>☒Image</p>
</blockquote></td>
<td><p>☐Less than 1 million images</p>
<p>☒1Million to1 billion images</p>
<ul>
<li><blockquote>
<p>More than 1 billion image</p>
</blockquote></li>
</ul></td>
<td><blockquote>
<p><em>Natural photographs, web and mobile UI screenshots, scientific
diagrams and charts, scanned documents and receipts, handwritten text
and math formulas, infographics, rendered LaTeX equations, synthetic
geometric shapes, abstract reasoning puzzles, desktop application
interfaces, map visualizations, hateful meme images, icons, tables,
arxiv paper figures, and annotated images with bounding boxes and point
labels.</em></p>
</blockquote></td>
</tr>
<tr class="odd">
<td><ul>
<li><blockquote>
<p>Audio (<em>Excluding audio that is part of video, as this should be
reported under the “video” modality instead. Furthermore, the Commission
understands the modality of ‘audio’ to include ‘speech’)</em></p>
</blockquote></li>
</ul></td>
<td><ul>
<li><blockquote>
<p>Less than 10 000 hours</p>
</blockquote></li>
<li><blockquote>
<p>10 000 to1 million hours</p>
</blockquote></li>
<li><blockquote>
<p>More than 1 million hours</p>
</blockquote></li>
</ul></td>
<td><blockquote>
<p><em>N/A</em></p>
</blockquote></td>
</tr>
<tr class="even">
<td><ul>
<li><blockquote>
<p>Video</p>
</blockquote></li>
</ul></td>
<td><ul>
<li><blockquote>
<p>Less than 10 000 hours</p>
</blockquote></li>
<li><blockquote>
<p>10 000 to1 million hours</p>
</blockquote></li>
<li><blockquote>
<p>More than 1 million hours</p>
</blockquote></li>
</ul></td>
<td><blockquote>
<p><em>N/A</em></p>
</blockquote></td>
</tr>
<tr class="odd">
<td><ul>
<li><blockquote>
<p>Other</p>
</blockquote></li>
</ul></td>
<td><blockquote>
<p><em>Specify the modality and for each one indicate approximate size
and unit of measurement</em></p>
</blockquote></td>
<td>N/A</td>
</tr>
</tbody>
</table>

**1.3.2 Latest date of data (acquisition/collection for model
training):**

<span class="mark">Feb 2026</span>

**1.3.3 Is data collection ongoing to update the model with new data
collection after deployment?**

No

**1.3.4 Date the training dataset was first used to train the model:**

Feb 2025

**1.3.5 Description of the linguistic characteristics of the overall
training data:**

The vision-language data used in the fine-tuning of this model is
primarily English.

**1.3.6 Other relevant characteristics of the overall training data:**

*N/A*

**1.3.7 Rationale or purpose of data selection:** High-quality
open-source datasets were selected to teach the model broad general
visual perception (caption, vqa, perception, OCR, logic, diagramqa) as
well as high-quality open-source datasets with math, STEM, and
computer-use for intended specialties in these domains.

## 2. List of data sources

### 2.1 Publicly available datasets

**2.1.1 Have you used publicly available datasets to train the model?**

Yes

**2.1.2 If yes, specify the modality(ies) of the content covered by the
datasets concerned:**

Text and Image

**2.1.3 List of large publicly available datasets:**

We used a variety of large, publicly available datasets including
Hugging Face and GitHub public repositories. More information on open
source training datasets is available in the associated [Technical
Report](https://www.microsoft.com/en-us/research/wp-content/uploads/2026/03/Phi-4-reasoning-vision-15B-Tech-Report.pdf?msockid=0261ed514fe0696c13f6fb8e4e6b68a3).

**2.1.4 General description of other publicly available datasets not
listed above:**

Additional vision-language datasets chosen to teach general vision
(caption, vqa, perception, OCR, logic, diagramqa) as well as math, STEM,
and computer-use capabilities.

### 2.2 Private non-publicly available datasets obtained from third parties

#### 2.2.1 Datasets commercially licensed by rightsholders or their representatives

**2.2.1.A Have you concluded transactional commercial licensing
agreement(s) with rightsholder(s) or with their representatives?** Yes

**2.2.1.B If yes, specify the modality(ies) of the content covered by
the datasets concerned:**

Text and Image

#### 2.2.2 Private datasets obtained from other third parties

**2.2.2.A Have you obtained private datasets from third parties that are
not licensed as described in Section 2.2.1, such as data obtained from
providers of private databases, or data intermediaries?** No

**2.2.2.B If yes, specify the modality(ies) of the content covered by
the datasets concerned:** N/A

**2.2.2.C If publicly known, list private datasets obtained from other
third parties:**

*N/A*

**2.2.2.D General description of non-publicly known private datasets
obtained from third parties:**

*N/A*

### 2.3 Data crawled and scraped from online sources

**2.3.1 Were crawlers used by the provider or on behalf of?**

*No*

**2.3.2 If yes, specify crawler name(s)/identifier(s):**

*N/A*

**2.3.3 Purposes of the crawler(s):**

*N/A*

**2.3.4 General description of crawler behavior:**

*N/A*

**2.3.5 Period of data collection:** MM/YYYY to MM/YYYY

*N/A*

**2.3.6 Comprehensive description of the type of content and online
sources crawled:** *N/A*

**2.3.7 Type of modality(ies) covered:**

*N/A*

**2.3.8 Summary of the most relevant domain names crawled:**

*N/A*

### 2.4 User data

**2.4.1 Was data from user interactions with the AI model (e.g. user
input and prompts) used to train the model?**

No

**2.4.2 Was data collected from user interactions with the provider’s
other services or products used to train the model?**

No

**2.4.3 If yes, provide a general description of the provider’s services
or products that were used to collect the user data:**

N/A

**2.4.4 Type of modality covered:**

N/A

### 2.5 Synthetic data

**2.5.1 Was synthetic AI-generated data created by the provider or on
their behalf to train the model?**

Yes

**2.5.2 If yes, modality of the synthetic data:**

Text

**2.5.3 If yes, specify the general-purpose AI model(s) used to generate
the synthetic data if available on the market:**

Open AI GPT-4o and o4-mini

**2.5.4 Information about other AI models, including model developer’s
own AI model(s) not available on the market, used to generate synthetic
data to train the model to which this Summary applies:**

N/A

<u>**2.5.5 Provide a description of the need or desired purpose for
using synthetic data for a model or system’s intended purpose:**</u>

Synthetic data consisted of high-quality captions and descriptions of
web images.

### 2.6 Other sources of data

**2.6.1 Was personal data used to train the model?** Microsoft follows
applicable laws and best practices pertaining to personal data.

**2.6.2 Have data sources other than those described in Sections 2.1 to
2.5 been used to train the model?**

Yes

**2.6.3 If yes, provide a narrative description of these data sources
and the data:**

Recordings of UI sessions capturing sequential screen states of desktop
navigation, web, or mobile interfaces, performing tasks such as
clicking, typing, scrolling, dragging, and switching between
applications, with each frame annotated with the corresponding action,
cursor position, and target element. This data was collected
specifically for model training.

## 3. Data processing aspects

### 3.1 Respect of reservation of rights from text and data mining exception or limitation

**3.1.1 Are you a Signatory to the Code of Practice for general- purpose
AI models that includes commitments to respect reservations of rights
from the TDM exception or limitation?**

Yes

**3.1.2 Describe the measures implemented before model training to
respect reservations of rights from the TDM exception or limitation
before and during data collection, including the opt-out protocols and
solutions honoured by the model developer or, as applicable, by third
parties from which datasets have been obtained:**

N/A

**<u>3.1.3 Does this dataset include any data protected by copyright,
trademark, or patent?</u>**

Microsoft follows applicable laws and best practices for processing data
protected by copyright, trademark, or patent.

### 3.2 Removal of illegal content

**3.2.1 General description of measures taken to avoid or remove illegal
content:**

Microsoft follows applicable laws and best practices to avoid or remove
illegal content.

### 3.3 Other information

**<u>3.3.1 Does the dataset include information about consumer groups
without revealing individual consumer identities?</u>**

Microsoft follows applicable laws and best practices for protecting
consumer identities.

<u>**3.3.2 Was the dataset cleaned or modified before model
training?**</u>

Yes