Discover What’s New in ChatGPT, Gemini, Claude, Copilot & Grok
Latest approved edition ·
THE EDITORIAL PROMISE
Official sources. Practical context. Reviewed before you read.
THE BRIEFING
What’s new. Why it matters.
All approved briefings · 22 reviewed stories
22 stories
ChatGPTOpenAIModels and availability
THE LEAD STORY
GPT-6.1 Terra starts rolling out in ChatGPT Work and Codex
OpenAI is introducing Terra in its Work and Codex experiences for eligible paid and education plans.
Rolling outPaid
Read more
What’s New?
The release notes describe support for visual reasoning, computer use and writing, with model selection becoming available as accounts receive the rollout.
Why It Matters
Teachers may encounter a new model option; this announcement does not establish its accuracy on chemistry tasks.
Availability
Rolling out to Plus, Pro, Business, Enterprise and Edu; access depends on plan, region and workspace settings. Overall plan usage limits are unchanged. UAE inference residency is unsupported.
OpenAI asks academic groups to help interpret AI research outputs
OpenAI announced research outputs in mathematics, computer science and physics alongside different approaches to expert review and explanation.
AnnouncedNot confirmed
Read more
What’s New?
Some artifacts are public, while release of physics materials is being left to an external collaboration. OpenAI describes the manuscripts as raw research data.
Why It Matters
The report illustrates the difference between producing a technical result and developing independently assessed scientific understanding. Claims remain attributed to OpenAI.
Availability
Some research artifacts are public; access to the underlying internal research models is not confirmed. This is not a new ChatGPT subscription feature.
CodeRabbit describes using GPT-6.1 Sol for security reviews
OpenAI published a case study of CodeRabbit using Sol to examine security risks spanning files and repositories.
AnnouncedNot confirmed
Read more
What’s New?
The report describes sandboxed verification for selected findings and proposed fixes submitted for human review.
Why It Matters
It is a third-party deployment example, not evidence that AI-generated software is secure or that classroom tools need no testing.
Availability
CodeRabbit product access, pricing and plan requirements are not confirmed in this source. These workflows are not announced as standard ChatGPT features.
OpenAI announces a forthcoming Google Cloud integration
OpenAI and Google Cloud announced plans to bring OpenAI models and Codex workflows into Google Cloud's Model Garden.
AnnouncedPaid
Read more
What’s New?
Organizations will be able to purchase the models through Google Cloud Marketplace; general availability is planned in coming weeks.
Why It Matters
Institutions already using Google Cloud may gain another procurement and deployment option.
Availability
Commercial purchases are planned through Google Cloud Marketplace. General availability is planned in coming weeks; a specific launch date is not confirmed.
OpenAI acquires Orchestra Bio to expand AI-assisted drug development
OpenAI says it has acquired Orchestra Bio and plans to combine the company’s drug-development coordination technology with its Rosalind tools for life-sciences research.
AnnouncedNot confirmed
Read more
What’s New?
The announced acquisition brings Orchestra Bio’s team and platform together with OpenAI’s Rosalind research tools. OpenAI describes a goal of helping scientists connect findings with studies, decisions, and next steps.
Why It Matters
Chemistry and biology students can see how AI may support the long, coordinated process of turning scientific findings into potential medicines, while human scientists remain responsible for experimental and clinical decisions.
Availability
Product access, pricing, and rollout details are not confirmed in this acquisition announcement.
ChatGPT can now hide selected saved conversations from regular history, pinned chats, and search, while allowing access with an account PIN or supported device authentication.
Rolling outFree & paid
Read more
What’s New?
The new locked-chats area protects eligible saved conversations and their chat-created files without deleting them. The feature follows existing retention and model-improvement settings.
Why It Matters
Students and educators gain another account-level way to reduce casual exposure of selected conversations on shared or visible devices, though locked chats do not replace school privacy policies or secure data-handling practices.
Availability
Rolling out gradually to Free, Go, Plus, and Pro users on the web, iOS, Android, macOS, and Windows.
xAI describes Grok Bot workflows for detecting and clearing support incidents
xAI published a case study describing two Grok Bot teammates: IncidentWatch looks for spikes in similar support tickets, while IncidentSweep prepares a reviewed batch of follow-up messages after an incident.
AnnouncedNot confirmed
Read more
What’s New?
The case study says xAI added two incident-focused AI teammates to its support workflow. A human support engineer investigates alerts and reviews a proposed message batch before approving any send.
Why It Matters
The example shows a practical pattern for using AI agents in monitored workflows: detect patterns, assemble evidence, pause for human review, and record the action.
Availability
Public access, plan eligibility, and pricing for these incident workflows are not confirmed by the case-study announcement.
Anthropic announced a fund linked to company equity to support public-benefit work. Its initial commitment remains subject to stockholder approval.
AnnouncedNot confirmed
Read more
What’s New?
The October 18 announcement describes a new fund and says more governance details will follow in coming months.
Why It Matters
This provides a concrete commitment to follow when discussing AI companies’ responsibilities and how benefits reach society.
Availability
Announced; the initial equity commitment is subject to stockholder approval. Grant eligibility, application timing and any new Claude access are not confirmed.
Background · Google announces Gemini 4 Argon for Pro and Ultra subscribers
Google’s October 13 release notes describe Gemini 4 Argon for complex tasks. This background entry introduces the platform; it is not a new announcement today.
Rolling outPaid
Read more
What’s New?
The release notes describe support for work such as analyzing complex reports and assembling presentations from multiple sources.
Why It Matters
Students can compare the documented tasks with their needs before checking subscription requirements.
Availability
Rolling out globally to Google AI Pro and Ultra subscribers in the Gemini app.
Background · Microsoft previews CRM context in Teams
Microsoft’s October 8 announcement describes bringing Dynamics 365 customer context into Teams. This is business software background, not a newly announced student feature.
TestingNot confirmed
Read more
What’s New?
The CRM experience in Teams is described as a public preview, connecting work conversations with customer information.
Why It Matters
It illustrates how AI assistance can depend on organizational data and permissions.
Availability
The CRM experience in Teams is in public preview. Exact licensing and student eligibility are not confirmed by this article; check Microsoft’s product requirements.
Background · xAI introduces Grok Voice 3.0 for its Voice Agent API
xAI’s October 14 announcement introduces Grok Voice 3.0, a speech-to-speech model offered through its Voice Agent API. This is dated background coverage.
Generally availablePaid
Read more
What’s New?
The announcement makes the voice model available through the API; it does not establish access for every consumer Grok plan.
Why It Matters
Developers can investigate spoken interfaces while checking API billing and deployment requirements.
Availability
Available through the paid Voice Agent API, subject to xAI’s API billing and access terms.
OpenAI announced presentation creation and shared editing in ChatGPT Space, including changes to text, visuals, and charts.
Rolling outPaid
Read more
What’s New?
Teams can edit a deck together, leave comments, and present or export it. Web export includes PowerPoint, PDF, and Google Slides.
Why It Matters
A group can develop and revise a presentation in one shared workspace.
Availability
Rolling out to eligible Pro, Business, and Enterprise users with Space access on web and desktop. The announcement does not promise access for every account.
Anthropic clarifies independent evaluators’ access and publication rights
Anthropic describes how METR evaluates its models and safety practices, with limits on sensitive information. An October 16 clarification updates the project terms.
AnnouncedNot confirmed
Read more
What’s New?
METR receives internal access and publication rights, subject to defined confidentiality restrictions.
Why It Matters
Readers can examine both the independence of external assessment and its limits.
Availability
This concerns evaluation arrangements, not a new user feature. Consumer plan access is not specified.
Tools from approved stories, with one original story for each announcement.
ChatGPT
GPT-Rosalind and Rosalind Workbench
The announcement connects AI tools with real scientific workflows in biology, chemistry, and drug development.
Try this: A class could compare an AI-generated research plan with the laboratory measurements, peer review, safety checks, and regulatory evidence needed before a treatment reaches patients.
Verify AI-generated chemistry explanations, calculations, citations, and structures before academic or laboratory use. Follow your course’s AI policy.
01
GPT-6.1 Terra · a suggested learning activity
Suggested use: try an ungraded explanation task and inspect every scientific claim. Independently verify AI chemistry answers and calculations before academic or laboratory use.
OpenAI research collaborations · a suggested learning activity
Suggested use: discuss what evidence would justify accepting an AI-generated scientific result. Independently verify AI chemistry answers and calculations before academic or laboratory use.
CodeRabbit security agent · a suggested learning activity
Suggested use: discuss why a chemistry simulator needs independent code and scientific review before release. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: draft a reusable chemistry-feedback checklist and review its permissions before trying it. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: discuss how a study app might route practice questions to topics; no integration is proposed for ChemHero. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: ask students to distinguish an adoption announcement from measured learning outcomes. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Grok Partner Network · a suggested learning activity
Suggested use: map which permissions a hypothetical campus integration would need. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: trace the difference between a measured cellular response, an AI-generated explanation, and a testable hypothesis. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: draft a fictional teaching-week checklist and identify which actions require instructor confirmation. Independently verify AI chemistry answers and calculations before academic or laboratory use.
OpenAI models and Codex · a suggested learning activity
Suggested use: compare a consumer chatbot subscription with an institution-managed cloud deployment. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: diagram the steps from an automated software warning through human review to a verified repair. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: compare a sponsored message with an independently supported scientific claim. Independently verify AI chemistry answers and calculations before academic or laboratory use.
GPT-Rosalind and Rosalind Workbench · a suggested learning activity
Suggested use: map the stages between a molecular discovery and a possible treatment, then discuss which stages require laboratory evidence and human oversight. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: compare locked chats, temporary chats, and ordinary chat history in a short digital-privacy lesson. Do not place confidential student or research data into an AI tool without institutional approval, and independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: ask students to diagram where human approval belongs in an AI-assisted laboratory alert or equipment-maintenance workflow. This is a discussion example, not tested chemistry performance; independently verify AI chemistry answers and calculations before academic or laboratory use.
Suggested use: discuss tests for an AI-rewritten calculation. Independently verify AI chemistry answers and calculations before academic or laboratory use.
Anthropic Public Benefit Fund · a suggested learning activity
Suggested classroom discussion: identify which commitments are conditional and what evidence would demonstrate progress. This announcement provides no confirmed classroom funding offer. AI chemistry answers and calculations need independent verification before academic or laboratory use.
Suggested classroom use: discuss how to verify a report summary against its sources. AI chemistry answers and calculations need independent verification before academic or laboratory use.
Microsoft Copilot and Dynamics 365 · a suggested learning activity
Suggested classroom use: compare a hypothetical campus service workflow with the business example. AI chemistry answers and calculations need independent verification before academic or laboratory use.
Suggested classroom use: diagram audio input, model processing and spoken output. This is not a validated laboratory tool. AI chemistry answers and calculations need independent verification before academic or laboratory use.
Suggested use: draft a short explanation of an acid–base titration, then check every calculation, figure, and chemical claim against course materials before presenting.
Claude safety evaluation · a suggested learning activity
Suggested discussion: identify what makes an evaluation independent and what evidence is missing. Verify AI-generated chemistry claims against reliable course sources.
