top of page
Search

AI Capability Releases: What's New in August 2026

Writer: claudinenm
claudinenm
Sep 7
6 min read

Cheaver's monthly guide to the latest updates from Claude, ChatGPT, Copilot, and Gemini



August 2026 delivered some of the most consequential AI platform updates of the year. Across all four major platforms, browser-based automation went mainstream, multi-model support became a standard enterprise expectation, and AI tools pushed into education, video, and voice at scale. Whether you are a business owner, a knowledge worker, or a developer, this month's releases have something that changes how you work. Here is everything you need to know.



Claude (Anthropic)

Mid-Conversation Tool Changes (Beta)

Claude Fable 5, Mythos 5, Opus 4.8, and Opus 5 now support adding or removing tools between turns of a conversation while preserving the prompt cache. This beta feature allows developers to dynamically adjust what Claude can access during a session without starting over, enabling more flexible and efficient agentic workflows.


Claude in Chrome Goes Generally Available

Claude in Chrome launched as generally available for every paid plan, including safer autonomous browser actions, prompt-injection defenses, and verified actions for reading, typing, clicking, and form-filling across connected web tools and legacy systems. Users can now delegate real browser-based tasks to Claude with appropriate safeguards in place.


Claude Developer Platform Playground

The Claude Console gained a fully featured Playground with complete Messages API support and runnable feature templates. Developers can now test every Messages API parameter interactively, with built-in templates that demonstrate features such as code execution and web search.


Claude Code Updates: Hooks, Visibility, and Security

Claude Code received new session hooks, richer usage and cost visibility, live Remote Control streaming for foreground subagents, and expanded CLI commands. The update also tightened security, permissions, and sandbox protections, making Claude Code more capable and more secure for development teams.


Managed Agent Controls for Enterprise

The Claude Developer Platform added managed agent controls covering session budgets, advisor models, inference geo pinning, and GitHub-hosted skills, giving enterprise teams stronger governance, deployment flexibility, and data residency options.


Claude for Teachers

Anthropic launched Claude for Teachers, a free Enterprise offering for K-12 schools and districts. The program provides educators with centrally managed access, single sign-on, role-based controls, and standards-aligned teaching tools, extending Claude's reach into education at no cost to qualifying institutions.




ChatGPT (OpenAI)

GPT-5.4 Mini Rolls Out to Free and Go Users

OpenAI rolled out GPT-5.4 mini broadly in ChatGPT, making it available to Free and Go tier users through the Thinking feature. The release expands access to more capable reasoning for users who previously had access only to older model versions.


Work Features Expand with Webhooks and Shared Tasks

ChatGPT Work gained webhook-triggered scheduled tasks, shared tasks across team members, and flexible limits for Free users. The Work browser was also expanded to complete tasks on signed-in websites, adding secure logins, password manager support, and confirmation steps before taking consequential actions on behalf of users.


Site Tools with WebMCP

OpenAI introduced Site tools with WebMCP, giving ChatGPT agents a direct way to work with websites through shared live pages and signed-in sessions. When available in the desktop app browser, ChatGPT Work and Codex can discover and use these tools automatically.


Enterprise Marketplace: GitHub Plugin Import

ChatGPT Enterprise and EDU administrators can now import and sync plugin marketplaces from GitHub, centrally distributing public or private repositories to workspace users. Automatic daily sync keeps plugins current without manual updates.


Atlas Deprecated, DALL-E GPT Retired

The Atlas browser agent was deprecated on August 9 as its capabilities were absorbed into ChatGPT and Codex. On August 30, the official DALL-E GPT was retired as image generation is integrated more directly into the core product. OpenAI also announced o3 will be retired August 26.



Microsoft Copilot

GPT-5 and Anthropic Model Support

Microsoft is rolling out GPT-5 in Copilot Chat and adding support for Anthropic models, including Claude. Microsoft 365 Copilot license holders receive access first. Word also gained model choice for document editing, letting users select from multiple AI providers.


Effort Level Controls

Copilot now lets users choose an effort level that balances quality, speed, and credit consumption: Light for simpler tasks, Medium as the default for everyday work, and High, Extra High, and Max for deeper analysis and complex reasoning.


Inline App Agents for Office Apps

Inline App Agents for Word, Excel, and PowerPoint allow users to add app-specific agents directly inside Copilot Chat to run actions without leaving the chat window, reducing context-switching during document work.


Copilot Notebooks Generally Available

Copilot Notebooks are now widely available for commercial and education customers and sync directly with OneNote, allowing users to organize AI-assisted research, notes, and drafts in a persistent structured format.


Python Analysis in Excel and Smarter Scheduling

Microsoft expanded AI-powered productivity across Excel, Outlook, Word, PowerPoint, OneNote, and Copilot Chat, with notable additions including Python analysis directly in Excel, richer email and meeting references, and smarter scheduling suggestions in Outlook.


Admin Analytics Export for Copilot Usage

IT administrators will soon be able to export anonymized organizational data related to Copilot and agent usage at a granular level, and can analyze GitHub Copilot AI Credit usage via the Consumption Dashboard.



Google Gemini

Gemini 3.6 Flash and 3.5 Flash-Lite

Google released Gemini 3.6 Flash with improved token efficiency and enhanced code and agentic planning at a lower price point than its predecessor. Gemini 3.5 Flash-Lite offers a low-latency, cost-effective option designed for high-volume automation, addressing developer feedback about output verbosity.


Gemini Omni 1.1 Flash Preview: High-Speed Multimodal

Gemini Omni 1.1 Flash is available in Preview, designed for video, image, and text tasks with high-speed generation and audio and video editing support. A new resolution parameter supports 360p, 720p, 1080p, and 4K outputs.


Gemini 3.5 Transcribe Generally Available

Gemini 3.5 Transcribe reached general availability with two dedicated speech-to-text models for high-accuracy transcription and low-latency live streaming. Supported features include automatic language detection, speaker diarization, word-level timestamps, custom vocabulary, and Live API streaming.


Student Tools and Free Google AI Plans

Google launched new student-focused features including a student hub, study notebooks, interactive visualizations, and Deep Research in Gemini Live. Eligible students can receive a free year of Google AI plans, signaling Google's push into the education market.


Ask Gemini in Google Chat

Ask Gemini is now integrated into Google Chat, providing Workspace-powered assistance for searching, drafting, catching up on conversations, managing tasks and events, and organizing work, all from within the Chat interface.


Google Meet Hardware: Notes Controller

Starting August 31, Google began rolling out the ability to start, stop, and manage the Take notes for me feature directly from the Google Meet Hardware touch controller for eligible meetings, bringing hands-free AI note management to conference rooms.



What Does This Mean for Users?

1. Browser Automation Is Now a Standard Feature

Both Claude in Chrome and ChatGPT Work's browser expansion went live or expanded meaningfully in August. For business users, this means AI tools can now complete real tasks inside web apps on their behalf, from filling in forms to updating records, with safeguards built in. You no longer need a developer to automate routine web workflows.


2. You Can Choose Your AI Model Inside Your Existing Tools

Microsoft Copilot's rollout of both GPT-5 and Claude model support in Word and Copilot Chat means business users can select the AI that works best for each task without leaving Microsoft 365. This marks a shift from single-vendor AI to a multi-model workplace environment.


3. AI Is Entering the Classroom in a Meaningful Way

Both Anthropic and Google launched education-specific offerings this month, with Claude for Teachers providing free enterprise-grade AI to K-12 schools and Gemini rolling out student study hubs with free annual access. If you work in education, now is the time to evaluate these tools.


4. Scheduled and Autonomous Tasks Are Becoming Routine

ChatGPT Work gained webhook-triggered scheduling, and Gemini deepened its Workspace integration, allowing AI to take actions on a timed or triggered basis without manual prompting. For small business owners, this means you can now set AI to handle recurring tasks automatically.


5. Voice and Video AI Has Reached Production Quality

Gemini 3.5 Transcribe's general availability, with speaker diarization, word timestamps, and live streaming, means AI transcription is no longer a premium add-on but a production-ready tool. Combined with Gemini Omni 1.1 Flash's 4K video output, the barrier to building voice- and video-enabled workflows has dropped significantly.


6. Enterprise Governance Tools Are Keeping Pace with AI Capabilities

Anthropic's session budgets and geo pinning for managed agents, combined with Microsoft's granular Copilot usage analytics, signal that enterprise AI governance is maturing. These controls give you meaningful oversight over how AI is being used, how much it costs, and where your data goes.


7. The AI Tool Ecosystem Is Consolidating

The deprecation of OpenAI's Atlas and DALL-E GPT, and the retirement of o3, reflect a deliberate consolidation of AI capabilities into core products. For users, this means your workflows will need fewer separate AI subscriptions as core platforms absorb more capabilities.



Stay in the Loop

This post is part of our monthly AI Capability Releases series. If you found it useful, subscribe to get the next edition delivered directly to your inbox. Have a question about how any of these features could work for your business? Drop a comment below or reach out directly.


Sources

Releasebot: Claude Developer Platform Updates, August 2026 — https://releasebot.io/updates/anthropic/claude-developer-platform

Releasebot: Claude Updates by Anthropic, August 2026 — https://releasebot.io/updates/anthropic/claude

OpenAI Help Center: ChatGPT Release Notes — https://help.openai.com/en/articles/6825453-chatgpt-release-notes

Releasebot: ChatGPT Updates by OpenAI, August 2026 — https://releasebot.io/updates/openai/chatgpt

Releasebot: Gemini Updates by Google, August 2026 — https://releasebot.io/updates/google/gemini

Google Workspace Updates: 2026 — https://workspaceupdates.googleblog.com/2026/

Comments

Rated 0 out of 5 stars.
No ratings yet
Add a rating

©2023 by Cheaver Consulting. All rights reserved.

  • Facebook
  • LinkedIn
bottom of page