LLM Changelog

Speaker recognition and attribution in Teams Rooms on Android
[Android]
Enhance your meetings with real-time speaker recognition and transcript attribution in Teams Rooms on Android. This feature identifies voices through cloud-enabled intelligent speakers and lets you securely enroll voices via Teams Settings-note that a Teams Rooms Pro license is required.
Learn more
.

Pro users are now able to use chat search connectors (in addition to deep research connectors) for the following integrations: Dropbox, Box, Google Drive (synced and non-synced), Microsoft OneDrive (Business), Microsoft SharePoint.
Note this feature is currently limited to users located outside of the EEA, Switzerland, and the UK.
Learn more aboutConnectors.

Plan ahead with scheduled actionsWhat:You can now schedule actions in the Gemini app so you can plan ahead for future and recurring tasks.Ask Gemini to help you get organized with a daily summary of your calendar, to-do's and important unread emails. Or get weekly creative ideas for a project and updates on topics you’re passionate about.You can schedule recurring actions, like finding local weekend music events every Friday, or one-time tasks, like asking for a recap of a keynote the next morning. Just tell Gemini what you need and when—it'll handle it for you.Scheduled actions are available to Google AI Pro and Ultra subscribers and qualifying Google Workspace business and education plans.Why:We want Gemini to be your personal, proactive and powerful AI assistant, and scheduled actions can help boost your productivity in new and exciting ways with proactive help that free up your time.

Plan ahead with scheduled actionsWhat:You can now schedule actions in the Gemini app so you can plan ahead for future and recurring tasks.Ask Gemini to help you get organised with a daily summary of your calendar, to-dos and important unread emails. Or get weekly creative ideas for a project and updates on topics that you're passionate about.You can schedule recurring actions, like finding local weekend music events every Friday, or one-time tasks, like asking for a recap of a keynote the next morning. Just tell Gemini what you need and when – it'll handle it for you.Scheduled actions are available to Google AI Pro and Ultra subscribers and qualifying Google Workspace business and education plans.Why:We want Gemini to be your personal, proactive and powerful AI assistant, and scheduled actions can help boost your productivity in new and exciting ways with proactive help that free up your time.

Capture meetings, brainstorms, or voice notes. Launching today for Pro, Enterprise, and Edu users. Previously launched for Team users on June 4, 2025.
ChatGPT will transcribe, summarize, and turn them into helpful outputs like follow-ups, plans, or even code.Available on the macOS desktop app only.
Learn more aboutChatGPT Record.

- Released `gemini-2.5-pro`, the stable version of our most powerful model, now with adaptive thinking. To learn more, see [Gemini 2.5 Pro](https://ai.google.dev/gemini-api/docs/models#gemini-2.5-pro) and [Thinking](https://ai.google.dev/gemini-api/docs/thinking). `gemini-2.5-pro-preview-05-06` will be redirected to `gemini-2.5-pro` on June 26, 2025.
- Released `gemini-2.5-flash`, our first stable 2.5 Flash model. To learn more, see [Gemini 2.5 Flash](https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash). `gemini-2.5-flash-preview-04-17` will be deprecated on July 15, 2025.
- Released `gemini-2.5-flash-lite-preview-06-17`, a low-cost, high-performance Gemini 2.5 model. To learn more, see [Gemini 2.5 Flash-Lite
Preview](https://ai.google.dev/gemini-api/docs/models#gemini-2.5-flash-lite).

Enterprise and Edu users can also now choose from the full set of ChatGPT models (GPT-4o, o3, o4-mini and more) when building Custom GPTs. This was launched to Plus, Pro and Team users earlier this month.
Key details:
GPTswithout Custom Actionscan use the model picker to select from all models available to the user.GPTswith Custom Actionscurrently support GPT-4o and 4.1Available onwebfor users on all paid plans (Plus, Pro Team, Enterprise, Edu).

We’ve upgraded ChatGPT search for all users to provide even more comprehensive, up-to-date responses. In testing, we found users preferred these search improvements over our previous search experience.
Improved quality
Smarter responses that are more intelligent, are better at understanding what you’re asking, and provide more comprehensive answers.Handles longer conversational contexts, allowing better intelligence in longer conversations.
Improved search capability and instruction following
More robust ability to follow instructions, especially in longer conversations, significantly reducing repetitive responses.Capability to run multiple searches automatically for complex or difficult questions.Search the web using an image you’ve uploaded.
Known limitations
Users may notice longer responses with this new search experience.In some cases, a “chain of thought” reasoning will show up unexpectedly for simple queries. A fix for this is rolling out to users shortly.ChatGPT may still make occasional mistakes - please double check responses.

Creators can now choose from the full set of ChatGPT models (GPT-4o, o3, o4-mini and more) when building Custom GPTs—making it easier to fine-tune performance for different tasks, industries, and workflows. Creators can also set a recommended model to guide users.
Key details:
GPTswithout Custom Actionscan use the model picker to select from all models available to the user.GPTswith Custom Actionscurrently support GPT-4o and 4.1Available onwebfor users onPlus, Pro and Teamplans.Enterprise and Edu rollout coming soon.

Starting today, we’re adding several updates to projects in ChatGPT to help you do more focused work. These updates are available for Plus, Pro, and Team users.
Deep research and voice mode supportImprovements to memory to reference past chats in a project*Sharing chats from projectsStarting a new project directly from a chatUpload files and access model selector on mobile
Learn moreabout projects.
*Memory improvements are available for Plus and Pro users.

Encrypt Copilot content with your own keys
[Web]
Microsoft Copilot Studio now allows you to use your own customer manager encryption keys (CMKs, hosted in Microsoft Azure Key Vault) to govern how Copilot Studio encrypts your copilot content.
By using customer managed encryption keys (CMKs), you can ensure that any data provided to your agent by your users, and the data you provide to Microsoft, is encrypted with your own keys.
You maintain control of your keys, providing you with further protection over the security of your data and ensuring you have control over how your data is stored at rest.
Key capabilities include hosting with Azure Key Vault to manage your keys, lifetimes, and rotation periods, encryption of all of your content, including copilot topics, settings and configurations, and conversation transcript data, rotation of CMKs, and, evocation of access, if necessary.
Learn more
.

Advanced email filtering in Copilot chat
[Windows]
Quickly surface exactly the emails you need—ask Microsoft 365 Copilot Chat for "last week's external emails," "threads I haven't replied to," "purple-category mail," or "summarize German emails" "emails where I'm on the To line"—and focus on what matters most.

Find emails awaiting your reply
[Windows]
Tell Microsoft 365 Copilot Chat "show me emails that I need to reply" and instantly see unread, read,
@mentioned
emails or emails with some question, task that you haven't answered—while hiding threads you've already closed—so you can clear your inbox with confidence.

Module UI refresh
[Windows, Web]
Copilot Chat is designed to provide a streamlined UI, making it easy to get started and achieve your goals quickly. It offers a helpful, understanding, and personalized experience, allowing you to search for past interactions, content, agents, or pages with ease.
Learn more
.

Gain DLP policy insights with Copilot
[Web]
Let Copilot instantly summarize Data Loss Prevention policies across locations, classifiers, and notifications. Use natural-language prompts to zoom into specific policies, spot gaps, and adjust settings faster—keeping your organization's data posture aligned without manual digging.
Learn more
.

Schedule a meeting from an email with Copilot
[Android, Windows, iOS, Mac, Web]
Turn any email thread into a meeting in one click. Copilot builds a complete invite for your review —title, agenda, attendee list, and a summary of the email conversation—plus attaches the original email thread so everyone is up to speed. Available in the new Outlook for Windows, web, Mac, and mobile.
Learn more
.

Improvements to the transcription experience in meetings
[Windows, Mac]
These updates enhance the transcription experience in meetings. When transcription, recording, or Copilot is enabled, users are prompted to choose the spoken language for accurate captions. Once transcription is running, only the organizer, co-organizers, and transcript initiator can change that language. A new settings page under Caption settings > Language settings > Meeting spoken language, along with a matching option under Transcript > Language settings, streamlines configuration. If someone speaks a language that doesn't match the selected one, The organizer/co-organizer and initiator receive a mismatch notification so they can adjust quickly.
Learn more
.

Microsoft Teams: Intelligent recap support for ad-hoc meetings and calls in GCC High
[Android, Windows, iOS, Mac, Web]
Intelligent meeting recap is now available in GCC High for impromptu calls and meetings, like those started from 'Meet now' and calls started from chat. You can easily browse the recording by speakers and topics, as well as access AI-generated notes, AI-generated tasks, and name mentions after the ad-hoc meeting ends. This capability is available for users with a Teams Premium or M365 Copilot license.

Advanced email filtering in Copilot chat
[Windows]
Quickly surface exactly the emails you need-ask Microsoft 365 Copilot Chat for "last week's external emails," "threads I haven't replied to," "purple-category mail," or "summarize German emails" "emails where I'm on the To line"-and focus on what matters most.

Find emails awaiting your reply
[Windows]
Tell Microsoft 365 Copilot Chat "show me emails that I need to reply" and instantly see unread, read,
@mentioned
emails or emails with some question, task that you haven't answered-while hiding threads you've already closed-so you can clear your inbox with confidence.

Like o1-pro, o3-pro is a version of our most intelligent model, o3, designed to think longer and provide the most reliable responses. Since the launch of o1-pro, users have favored this model for domains such as math, science, and coding—areas where o3-pro continues to excel, as shown in academic evaluations. Like o3, o3-pro has access to tools that make ChatGPT useful—it can search the web, analyze files, reason about visual inputs, use Python, personalize responses using memory, and more. Because o3-pro has access to tools, responses typically take longer than o1-pro to complete. We recommend using it for challenging questions where reliability matters more than speed, and waiting a few minutes is worth the tradeoff.
In expert evaluations, reviewers consistently prefer o3-pro over o3 in every tested category and especially in key domains like science, education, programming, business, and writing help. Reviewers also rated o3-pro consistently higher for clarity, comprehensiveness, instruction-following, and accuracy.
Academic evaluations show that o3-pro consistently outperforms both o1-pro and o3.
To assess the key strength of o3-pro, we once again use our rigorous "4/4 reliability" evaluation, where a model is considered successful only if it correctly answers a question in all four attempts, not just one:
o3-pro is available in the model picker for Pro and Team users starting today, replacing o1-pro. Enterprise and Edu users will get access the week after.
As o3-pro uses the same underlying model as o3, full safety details can be found in theo3 system card.
Limitations
At the moment, temporary chats are disabled for o3-pro as we resolve a technical issue.
Image generation is not supported within o3-pro—please use GPT-4o, OpenAI o3, or OpenAI o4-mini to generate images.
Canvas is also currently not supported within o3-pro.

Gain DLP policy insights with Copilot
[Web]
Let Copilot instantly summarize Data Loss Prevention policies across locations, classifiers, and notifications. Use natural-language prompts to zoom into specific policies, spot gaps, and adjust settings faster-keeping your organization's data posture aligned without manual digging.
Learn more
.

Schedule a meeting from an email with Copilot
[Android, Windows, iOS, Mac, Web]
Turn any email thread into a meeting in one click. Copilot builds a complete invite for your review -title, agenda, attendee list, and a summary of the email conversation-plus attaches the original email thread so everyone is up to speed. Available in the new Outlook for Windows, web, Mac, and mobile.
Learn more
.

We're upgrading Advanced Voice in ChatGPT for paid users with significant enhancements in intonation and naturalness, making interactions feel more fluid and human-like. When we first launched Advanced Voice, it represented a leap forward in AI speech—now, it speaks even more naturally, with subtler intonation, realistic cadence (including pauses and emphases), and more on-point expressiveness for certain emotions including empathy, sarcasm, and more.
Voice also now offers intuitive and effective language translation. Just ask Voice to translate between languages, and it will continue translating throughout your conversation until you tell it to stop or switch. It’s ready to translate whenever you need it—whether you're asking for directions in Italy or chatting with a colleague from the Tokyo office. For example, at a restaurant in Brazil, Voice can translate your English sentences into Portuguese, and the waiter’s Portuguese responses back into English—making conversations effortless, no matter where you are or who you're speaking with.
This upgrade to Advanced Voice is available for all paid users across markets and platforms—just tap the Voice icon in the message composer to get started.
This update is in addition to improvements we made earlier this year to ensure fewer interruptions and improved accents.
Known Limitations
In testing, we've observed that this update may occasionally cause minor decreases in audio quality, including unexpected variations in tone and pitch. These issues are more noticeable with certain voice options. We expect to improve audio consistency over time.
Additionally, rare hallucinations in Voice Mode persist with this update, resulting in unintended sounds resembling ads, gibberish, or background music. We are actively investigating these issues and working toward a solution.

- Released `gemini-2.5-pro-preview-06-05`, a new version of our most powerful model, now with adaptive thinking. To learn more, see [Gemini 2.5 Pro Preview](https://ai.google.dev/gemini-api/docs/models#gemini-2.5-pro-preview-06-05) and [Thinking](https://ai.google.dev/gemini-api/docs/thinking). `gemini-2.5-pro-preview-05-06` will be redirected to `gemini-2.5-pro` on June 26, 2025.

ChatGPT Team, Enterprise, and Edu customers globally can use connectors in deep research, as well as Pro and Plus users (excluding users in Switzerland, EEA, and the UK) to generate long-form, cited responses that include your company’s internal tools.
Supported connectors: Google Drive, SharePoint, Dropbox, Box, Outlook, Gmail, Google Calendar, Linear, GitHub, HubSpot, and TeamsCombines internal + web sources for synthesis
Learn more aboutConnectors in ChatGPT.