Dateline: August 21, 2026 | Next update: August 28, 2026
A two-week window with no new model launch but several platform-level moves that affect every developer and enterprise. The headline pricing story: on August 10, Anthropic made Claude Sonnet 5's $2/$10 per MTok pricing permanent, cancelling the September 1 increase to $3/$15 that teams had been budgeting for. Three APIs graduated from beta to GA simultaneously — Files API, Agent Skills, and Admin API user management. The Workbench and three experimental prompt endpoints retired on August 17 with no recovery path for missed exports. Managed Agents gained session budgets, an advisor model parameter, inference geo-pinning, and GitHub-hosted skills. Claude Code shipped two broad updates covering its Concise output style, cross-session idle notifications, macOS sandbox hardening, and a range of security and reliability fixes. Claude Academy launched. Anthropic is reportedly in talks to acquire AI infrastructure startup Decart for approximately $6 billion. On the OpenAI side, Ultrafast previewed as a GPT-5.6 Sol service tier running up to 14× Standard speed; ChatGPT Business gained Premium seats; GPT-5.6-Cyber and expanded Daybreak access launched for approved defenders; and internal evaluations that could not rule out the upcoming Astra model reaching the Critical cyber threshold led OpenAI to pause frontier reinforcement-learning training and expand monitoring. o3 leaves ChatGPT on August 26 and the official DALL·E GPT retires August 30. Google celebrated the Gemini app passing one billion active users and unveiled Pixel 11 AI features at Made by Google 2026.
Claude / Anthropic
✅ Claude Sonnet 5 — $2/$10 pricing made permanent
On August 10, Anthropic made Claude Sonnet 5's $2/$10 per MTok pricing permanent, cancelling the scheduled September 1 step-up to $3/$15. This is not a price cut — nobody's bill goes down. It is a withdrawn 50% increase that teams had been budgeting for. Any team that built Q4 projections assuming September rates can stand down. Competitive pressure from OpenAI GPT-5.6 Sol is widely cited as the driver.
Sonnet 5 permanent pricing confirmed August 10: $2 per million input tokens, $10 per million output tokens — this is now the standard price, not a promotional rate. The September 1, 2026 step-up to $3/$15 will not occur, and Anthropic's official launch post has been updated to reflect this. Two cost mechanisms remain unchanged: prompt cache hits bill at 10% of the standard input rate ($0.20/MTok); the Batch API cuts both input and output by 50% ($1/$5 per MTok). Important tokenizer reminder: Sonnet 5 uses a newer tokenizer that maps the same text to 1.0–1.35x more tokens than Sonnet 4.6. At $2/$10, Sonnet 5 is still cheaper than Sonnet 4.6 ($3/$15) for most workloads, but the effective gap is narrower than the headline numbers suggest on prose-heavy content.
Standard price: $2/$10 per MTok (permanent from August 10) | Previously scheduled from Sept 1: $3/$15 (cancelled) | Cache hits: $0.20 per MTok input | Batch API: $1/$5 per MTok | Tokenizer: 1.0–1.35x more tokens than Sonnet 4.6 for the same text | Model string: claude-sonnet-5 | Subscription: unchanged (plan limits, not per-token billing)
Best for: Every team running Sonnet 5 workloads: update your cost models to a permanent $2/$10. If you budgeted for September at $3/$15, that budget is now available for capacity.
★ API — Files API, Agent Skills and Admin API user management reach GA
Three significant APIs graduated from beta to general availability in the same period — an unusually broad simultaneous GA push. All three had been in beta for months; the simultaneous graduation signals Anthropic is stabilising the developer platform ahead of its IPO filing.
Files API is now GA. Requests to /v1/files endpoints and Messages API requests referencing uploaded files no longer require the files-api-2025-04-14 beta header. The GA response format adds file expiration (set expires_in_seconds on upload; file objects now report expires_at), page and next_page pagination, and an ids[] filter on list requests. Storage is 1 TB per organisation with a 500 requests per minute rate limit. Requests still sending the beta header continue to work and receive the previous response format. Agent Skills API is now GA: /v1/skills endpoints and Messages API requests loading Skills via the container parameter no longer require the skills-2025-10-02 beta header. Admin API user management is now GA for Claude Enterprise — group and custom-role endpoints no longer require the anthropic-beta: ce-user-management-2026-07-13 header.
Files API GA: no header; expires_in_seconds on upload; expires_at on file object; page/next_page pagination; ids[] filter; 1 TB/org; 500 req/min | Agent Skills GA: no header on /v1/skills or the container parameter | Admin API user management GA: no header on group and custom-role endpoints | All three: legacy beta headers still accepted for backward compatibility
Best for: API developers — remove the three beta headers from your code. No breaking changes, since the headers are still accepted, so there is no urgent migration. Update SDKs to get the GA response formats for the Files API.
★ Claude Managed Agents — session budgets, advisor model, geo-pinning, GitHub skills
Managed Agents received four new capabilities this period, each addressing a distinct production-readiness gap. Together they give enterprise teams cost governance, quality control, data-residency compliance, and version-controlled skill deployment in a single update.
Session budgets: set a hard spend cap when creating a session by passing a budget object with type 'limit' and max_list_cost in US cents. The cap counts model token cost at public list price, summed across all threads including subagents; when it is reached, no new subagents spawn and running background agents halt. Advisor model parameter: specify which Claude model acts as the advisor in an agent session independently of the executor model — previously this defaulted to the session model. Inference geo-pinning: set an inference region on an agent to pin where inference runs, enabling data residency compliance without routing all traffic through a single region. GitHub-hosted skills: point an agent at a GitHub repository to load skills directly, with version pinning via git ref — skills live in your own repository, follow your own review process, and update on push without any Anthropic-side configuration change. Web access controls: set allowed_domains or blocked_domains on web_search and web_fetch. A redesigned Console session viewer adds per-turn model and effort, tool call details, and subagent thread tracing.
Session budgets: budget.type 'limit', budget.max_list_cost.amount (string, US cents, no float), budget.max_list_cost.currency 'USD' | Advisor model: advisor_model on session or agent config | Geo-pinning: inference_region on agent | GitHub skills: github_ref on skill config | Web controls: allowed_domains / blocked_domains on tool config; web_fetch max_content_tokens; web_search user_location | All under managed-agents-2026-04-01 | SDK minimum for budgets: anthropic>=0.121.0
Best for: Enterprise teams needing cost guardrails on autonomous agents (budgets), data residency compliance (geo-pinning), version-controlled skill deployment (GitHub skills), and fine-grained web access control (domain allow/block lists).
★ API — workspace ID response header and Rate Limits API
Two small but operationally useful API additions shipped this period. Both address longstanding observability gaps for teams running multi-workspace or multi-team Claude deployments.
anthropic-workspace-id response header: the Claude API now returns this header on every response, carrying the wrkspc_-prefixed ID of the workspace that the request's API key or access token resolved to — including the Default Workspace. Useful for debugging routing issues and attributing spend across workspaces without inspecting the key itself. Rate Limits API: administrators can now programmatically query the rate limits configured for their organisation and workspaces via a new endpoint, rather than reading them only from the Console UI. This enables automated alerting and limit-aware request scheduling.
Header: anthropic-workspace-id on every API response | Value: wrkspc_-prefixed workspace ID | Rate Limits API: GET /v1/rate_limits (admin scope required) | Both available to all API users with appropriate credentials
Best for: Platform teams attributing API spend across workspaces, and developers building rate-limit-aware scheduling or alerting systems.
⚠ Workbench retired August 17 — no recovery path
The legacy Claude API Workbench at platform.claude.com/workbench was retired on August 17, 2026. Any organisation that stored prompts, variables, or evaluation configurations there and did not export them before that date has permanently lost access to those assets — Anthropic has confirmed no recovery path exists. The three experimental prompt-engineering API endpoints retired simultaneously.
Retired August 17: the legacy Workbench, /v1/experimental/generate_prompt, /v1/experimental/improve_prompt, and /v1/experimental/templatize_prompt. All three endpoints now return errors. A new stateless Workbench (the Playground at platform.claude.com) replaces the legacy version — the key difference is that it stores no data server-side; drafts exist in the browser only. For teams who missed the export window, Anthropic has confirmed no data recovery is available. The three endpoint functions can be replicated using a standard Messages API meta-prompt approach; see Claude Docs for migration guidance.
Best for: If your automation still calls the three retired endpoints, it is now broken — migrate to the Messages API meta-prompt approach immediately.
★ Claude Academy — a learning hub for AI fluency
Anthropic launched Claude Academy, a structured learning hub for safe, effective AI use. It is aimed at the large population of people who visit Anthropic's site to learn how to use AI but had no formal curriculum to follow. The content spans beginner to advanced, covering delegation, verification, and critical evaluation of AI outputs alongside practical skill-building.
Claude Academy is live at anthropic.com. Courses and tutorials cover AI fluency from beginner to advanced levels, with badges on course completion and personalised learning recommendations based on role and goals. Content was developed by Anthropic's in-house education team, emphasising practical skills: how to delegate effectively to Claude, how to verify outputs, and how to build healthy habits around AI use. Free for all users. Role-based certifications for Claude deployment professionals were announced alongside the Academy — Anthropic's first formal credential for the forward-deployed engineer and agent manager roles now appearing in enterprise hiring.
URL: anthropic.com/academy | Cost: free | Audience: all Claude users, from first-time to advanced | Format: courses, tutorials, completion badges, personalised recommendations | Certifications: role-based, for Claude deployment professionals | Developed by: Anthropic's in-house education team
Best for: Teams onboarding new Claude users at scale, individuals building AI fluency, enterprise L&D teams looking for structured AI training, and forward-deployed engineers seeking formal certification.
★ Claude Code — Concise output style, idle notifications, macOS sandbox hardening
Claude Code shipped two broad point releases across the two-week window. The most user-visible addition is the Concise output style, which trains Claude to lead with results and skip preamble. The most security-significant is macOS sandbox hardening.
Concise output style: a new built-in style available under Output style in /config, where Claude leads with results and skips preamble and narration while doing the work just as thoroughly. Cross-session idle notifications: Claude now notifies you when a background session has been idle and waiting for input, so sessions no longer stall silently. macOS sandbox hardening reduces the blast radius of a compromised tool call. A new default model setting lets admins in managed environments set an org-level default that applies to all sessions, shown in /model as 'Org default'. Prompt caching is fixed for custom gateways: sessions using an LLM gateway or custom ANTHROPIC_BASE_URL now correctly apply caching, fixing a regression that had disabled it silently. Remote Control effort changes made on a phone or claude.ai/code now propagate to terminal and Desktop sessions in real time. GitLab MR URL support was added to the --worktree flag and agents view. Linux memory cgroup support for the Bash tool prevents runaway builds stalling a session. Session titles now render as short, specific names rather than sentences.
Concise style: Output style → /config | Cross-session idle notifications: automatic | Org default model: claude.ai admin → org settings | Prompt caching fix: LLM gateway / custom ANTHROPIC_BASE_URL now caches correctly | Remote Control effort: propagates from mobile/web to terminal/Desktop | GitLab MR: --worktree flag + agents view (!N format) | CLAUDE_CODE_TOOL_MEMORY_LIMIT: Linux Bash memory cgroup | forward_user_identity: opt-in on Anthropic gateway upstreams for per-user spend attribution | keybindingFlavor: set to 'readline' for Ctrl+W to delete to previous whitespace
Best for: Experienced Claude Code users who want faster output (Concise style), teams running background agents (idle notifications), macOS users on auto mode (sandbox hardening), and gateway users who hit the caching regression.
⚠ Anthropic reportedly in talks to acquire Decart (~$6B)
Anthropic is reportedly in advanced talks to acquire Decart, an AI infrastructure startup specialising in real-time inference optimisation, for approximately $6 billion. The deal, if it closes, is aimed squarely at reducing the inference cost that underlies Anthropic's pricing decisions. No closing date has been announced and the deal has not been officially confirmed.
Decart acquisition talks reported August 13 by multiple outlets, at a reported valuation of approximately $6 billion. Strategic rationale: Decart specialises in real-time inference optimisation — squarely the cost driver behind the Sonnet 5 pricing decision. Anthropic has neither confirmed nor denied. Strategic context: this pairs with the in-house silicon team confirmed August 5 and compute deals with SpaceX (300MW), Amazon (5GW) and Google (5GW). Anthropic filed a confidential S-1 at approximately $350B valuation on June 1.
Best for: Informational — no action required. A long-term signal for inference cost trajectory and Anthropic's vertical integration strategy.
Plans and pricing
The most significant pricing change of the period is the permanent Sonnet 5 rate. The only remaining time-sensitive limit boost is Claude Code's 50% higher weekly limits, which expired August 19.
Sonnet 5: $2/$10 per MTok (permanent; September 1 increase cancelled) | Batch API Sonnet 5: $1/$5 per MTok | Fable 5 API: $10/$50 per MTok, Batch $5/$25 | Opus 5: $5/$25 standard, $10/$50 fast mode | Opus 4.8: $5/$25 (available) | Haiku 4.5: low-cost tier | Claude Code limits: +50% through August 19 (now expired) | Fable 5 subscription: Max/Team Premium 50% of weekly limits included; Pro/Team Standard usage credits only | No model retirements currently scheduled | Workbench retired August 17
Best for: Sonnet 5 budgets can now be set permanently at $2/$10 — remove any September rate-increase assumptions from Q4 plans.
ChatGPT / OpenAI
★ GPT-5.6 Sol Ultrafast — up to 14× faster frontier inference
OpenAI previewed Ultrafast, a new API service tier designed to run GPT-5.6 Sol at dramatically higher inference speeds — up to 14 times the speed of Standard processing and as many as 750 output tokens per second, while retaining the intelligence of the full model.
The significance is less about benchmark capability than latency. Developers have traditionally had to choose smaller models when they need near-real-time responses. Ultrafast removes some of that trade-off, making the flagship reasoning model practical for latency-sensitive coding, voice, financial analysis and interactive agents.
Ultrafast launches first through the OpenAI API and is powered by Cerebras infrastructure. It is a distinct processing tier rather than a separate model: developers still use GPT-5.6 Sol but request substantially faster inference. OpenAI describes this as an early preview rather than a universal replacement for Standard processing.
Model: GPT-5.6 Sol | Service tier: Ultrafast | Platform: OpenAI API | Maximum stated speed-up: up to 14× Standard | Maximum stated generation speed: up to 750 output tokens/sec | Infrastructure partner: Cerebras | Model capability: same GPT-5.6 Sol | Status: early preview
Best for: Developers building real-time or latency-sensitive applications that need GPT-5.6 Sol-level intelligence without conventional frontier-model response times.
★ ChatGPT Business — Premium seats with 5× more usage
OpenAI announced a new Premium seat for ChatGPT Business, aimed at heavy users who repeatedly hit the capacity of a Standard seat. It provides five times more usage and removes the five-hour usage limit while keeping users inside the same centrally administered workspace.
Premium costs $125 per user per month billed monthly, or $100 per user per month billed annually. Standard seats remain $25 monthly or $20 annually. Crucially, organisations do not have to upgrade the whole workspace: Standard and Premium seats can coexist, and administrators can assign or reassign the higher-capacity seats to employees whose workloads justify them. OpenAI also offered the first 10,000 eligible Business customers who joined the waitlist by August 20 up to $500 in promotional workspace credits. That window has now closed.
Premium monthly: $125/user/month | Premium annual: $100/user/month | Standard monthly: $25/user/month | Standard annual: $20/user/month | Included capacity: 5× Standard | Five-hour limit: removed on Premium | Workspace: Standard + Premium seats can be mixed | Promotion: up to $500 workspace credits; waitlist deadline August 20
Best for: ChatGPT Business organisations with a smaller group of extremely heavy Work, Codex or agentic users. Upgrade those users rather than paying Premium pricing across the whole workforce.
★ GPT-5.6-Cyber and expanded Daybreak access
OpenAI substantially expanded its cybersecurity offering with GPT-5.6-Cyber, a cybersecurity-specific model, and two controlled Daybreak access levels. The move follows OpenAI's conclusion that frontier models are approaching capabilities that could materially change both offensive and defensive cybersecurity.
Daybreak Blue provides approved defenders access to frontier general-purpose models including GPT-5.6 Sol, with safeguards designed for authorised defensive work — vulnerability discovery, secure-code review, malware analysis, incident response and patch validation. More specialised cyber work can receive tighter controlled access through the Daybreak programme. The broader strategy is to give trusted defenders access to advanced cyber capabilities before similar capabilities become widely available to attackers, treating access controls, identity, scope, logging, monitoring and human oversight as part of the product. On August 11, Daybreak extended into AWS: both Blue and Red access levels are now available through Amazon Bedrock, letting organisations standardised on AWS bring these models into existing cloud security architectures.
New model: GPT-5.6-Cyber | Programme: OpenAI Daybreak | Blue: frontier general-purpose models including GPT-5.6 Sol with defensive safeguards | Workflows: vulnerability discovery, secure-code review, malware analysis, incident response, patch validation | AWS: Daybreak Blue and Red via Amazon Bedrock from August 11 | Governance: identity verification, defined testing scope, monitoring, logging, human oversight
Best for: Cybersecurity teams, incident-response organisations and enterprises that need frontier defensive capabilities under controlled governance — particularly AWS-heavy estates.
★ ChatGPT for Teens — dedicated under-18 experience
OpenAI launched ChatGPT for Teens, a dedicated experience that automatically applies stronger protections when a user states that they are 13–17 or OpenAI's systems estimate that they are under 18. The product combines educational features with more restrictive model behaviour in areas where younger users face greater risks.
The learning side combines Study Mode with quizzes, Learning Visualizations and a new Study Hours setting that allows teens or parents to specify periods when Study Mode should be enabled by default. Responsible-homework reminders recognise when a user appears to be trying to bypass the learning process and redirect them towards guided problem solving. Teen accounts receive stronger protections around self-harm, violence, eating disorders, dangerous activities and explicit material. OpenAI also explicitly instructs the teen experience not to encourage emotional dependence, use romantic language, or imply that ChatGPT has feelings or consciousness. Additional features include break reminders, sensitive-image warnings and teen-specific onboarding. Parents with linked accounts retain controls including Quiet Hours and limited safety notifications.
Age scope: 13–17 or system-estimated under 18 | Assignment: automatic when age criteria are met | Learning: Study Mode, quizzes, Learning Visualizations, Study Hours, responsible-homework reminders | Safety: stronger protections for self-harm, eating disorders, violence, dangerous activities and sexual/graphic content | Relational safeguards: no romantic language, no emotional-dependence encouragement, no claims of consciousness | Parent controls: Quiet Hours, selected settings, limited high-risk safety notifications
Best for: Families, educators and policymakers tracking how generative AI products are adapting for minors. For teen users, the key change is that safety settings are default rather than optional.
★ ChatGPT — interactive quizzes, editable Project memory, Google Drive in Library
ChatGPT gained interactive quizzes that let users answer questions directly inside a conversation rather than asking ChatGPT to generate a static quiz. Projects also became more flexible, and Google Drive became directly browsable from Library.
Eligible unshared Projects can now switch between default memory and project-only memory without creating a new Project. Project-only memory keeps context contained: ChatGPT can use conversations from the Project but does not reference memories or chats outside it, and Project information is not added to memory used elsewhere. Shared Projects remain project-only and cannot be switched; ChatGPT Work is unavailable inside project-only Projects. Separately, users who connect the Google Drive plugin can now browse Drive files and folders directly from Library, including items shared with them, and pull them into a conversation through the composer or @ mentions without re-uploading. Google Docs, Sheets and Slides can remain open alongside the conversation. Eligible paid users can also receive homepage suggestions based on conversation history and connected tools; Free and Go users gained the ability to select Think on the web.
Quizzes: consumer + Edu, web/mobile | Project memory: existing eligible unshared Projects can switch default ↔ project-only | Shared Projects: remain project-only | Work: unavailable in project-only Projects | Drive: files, folders and directly shared items via Library, composer or @ mention; re-upload not required | Homepage suggestions: eligible paid users | Think: Free and Go on web
Best for: Students using ChatGPT for active learning, professionals needing separation between Project-specific context and general memory, and Google Workspace users doing document-heavy research.
★ ChatGPT Business — Computer History on macOS, restaurant reservations
OpenAI introduced Computer History, an optional macOS feature that allows ChatGPT and Codex to use context from selected applications and websites. Rather than capturing screenshots or recordings, it records interaction events that can later help ChatGPT understand what the user was working on. Separately, ChatGPT gained real-time restaurant reservation search.
Computer History is off by default. A Business workspace administrator must first enable access, after which individual members decide whether to opt in. Users can pause collection, choose which applications and sites are included, and inspect or delete their timeline data; private browsing is excluded. At launch it is not available in the EEA, United Kingdom or Switzerland — a restriction particularly relevant for European enterprise deployments. Restaurant reservations: ChatGPT can now search real-time availability inside a conversation via OpenTable (global), Resy (US) and Yelp (US and Canada). Users provide a location, date, time, party size and preferences, and ChatGPT returns available slots rather than simply recommending restaurants. Rolling out across all ChatGPT plans on mobile, web and desktop; ChatGPT Work is excluded.
Computer History — platform: ChatGPT macOS | Plan: Business | Default: off | Admin must enable, user must opt in | Captured: interaction events | Not captured: screenshots, screen recordings, microphone or system audio | Private browsing excluded | Initial exclusions: EEA, UK, Switzerland | Reservations — partners: OpenTable (global), Resy (US), Yelp (US + Canada); all plans; Work excluded
Best for: Business users who want ChatGPT to retain desktop context across applications — European organisations should note the EEA/UK/Switzerland exclusion at launch.
⚠ Enterprise connected-app sync — individual connections retired
OpenAI retired individually authorised sync connections for connected enterprise apps. New individual-user sync connections stopped being available on August 10, and existing connections were disabled on August 14, with deletion of associated synced data beginning at that point.
The change does not eliminate enterprise sync — it moves organisations towards administrator-managed connections. Google Drive can use the Google Drive plugin and, where indexed knowledge is needed, domain-wide delegated admin sync. SharePoint follows a similar administrator-managed model, while GitHub moves to the non-synchronised GitHub plugin. Existing individual sync connections should no longer be treated as active; enterprise administrators who relied on them need to ensure replacement plugins or administrator-managed sync are configured. OpenAI said replacement plugin availability for GitLab Issues and Azure Boards would be communicated separately.
Best for: Enterprise and Edu admins — verify that important knowledge sources have moved from individual-user sync to supported administrator-managed integrations.
⚠ Astra may reach Critical cyber capability — OpenAI slows frontier training
The most consequential research and safety development of the period concerns Astra, an upcoming OpenAI model. Internal evaluations led OpenAI to conclude on August 7 that it could no longer rule out Astra reaching the Critical cybersecurity capability threshold under its Preparedness Framework.
On August 18, OpenAI disclosed that this assessment, together with lessons from the OpenAI–Hugging Face security incident, had already affected its model-development schedule. The company temporarily slowed frontier scaling, including a two-week pause in reinforcement-learning training for models intended for deployment; its largest planned frontier RL run remained on hold at publication. OpenAI now requires its strictest security environment for Astra and cyber-model workloads. Monitoring has been expanded to all RL training and evaluations involving tools for models at Sol capability or above, as well as all Astra inference that uses tools. The monitoring stack uses token-level activation classifiers and increasingly capable automated investigators to detect unauthorised access, data theft, destructive behaviour and attempts to defeat safeguards, with a target alert time within 30 minutes. OpenAI estimates the monitoring overhead at roughly 20% of the inference compute being monitored.
Best for: AI safety researchers, cybersecurity leaders and organisations tracking frontier-model timelines. The important implication is that capability growth is now materially changing OpenAI's internal training and deployment process.
⚠ The Defender's Window — OpenAI urges rapid AI-assisted security adoption
OpenAI published a broader warning that organisations have a limited period in which to strengthen their defences before increasingly capable AI systems make automated exploitation substantially easier. The company argues that the same capabilities can advantage defenders — but only if security teams deploy them quickly.
The recommendation is not simply to purchase an AI security tool. OpenAI argues for improving foundational controls while gradually introducing AI into vulnerability discovery, code review, alert triage, incident response and remediation, with autonomy expanding as organisations gain confidence. Recommended progression: foundational controls → AI-assisted code and security review → advisory scanning → live alert triage → narrowly scoped automated remediation. The warning follows OpenAI's assessment that it underestimated the real-world cyber capability demonstrated during the OpenAI–Hugging Face incident.
Best for: CISOs and enterprise security teams. The recommendation is to begin deploying AI-assisted defensive workflows now rather than waiting for autonomous offensive capability to become commonplace.
⚠ o3 retires from ChatGPT August 26 — DALL·E GPT August 30
OpenAI o3 is five days from its scheduled retirement in ChatGPT, following the 90-day sunset announced in May. The official DALL·E GPT retires four days after that.
The o3 retirement applies to ChatGPT only — OpenAI has not announced a corresponding API retirement. Users who still rely on o3-specific prompts or workflows should test them against GPT-5.6 reasoning models before August 26. The DALL·E GPT retirement does not remove image generation from ChatGPT: users should move to ChatGPT Images, and user-created GPTs with image generation enabled are unaffected. OpenAI recommends saving any images you want to retain before August 30.
Best for: Paid ChatGPT users still selecting o3 manually — move repeatable workflows to GPT-5.6 before August 26, and preserve anything from DALL·E GPT before August 30.
Plans and pricing
The major commercial change in this two-week window is ChatGPT Business Premium. GPT-5.6 model API list pricing did not change; the major API development is instead the Ultrafast preview, which creates a new performance tier where latency matters.
ChatGPT Business Standard: $20/user/month annual, $25 monthly | Premium: $100/user/month annual, $125 monthly | Premium capacity: 5× Standard | Five-hour usage limit: removed on Premium | Premium promotion: waitlist credit offer ended August 20 | GPT-5.6 API list-price change: none this period | Ultrafast: new GPT-5.6 Sol high-speed API tier | o3 ChatGPT retirement: August 26 | DALL·E GPT retirement: August 30
Best for: Business customers should identify the small number of employees whose usage justifies Premium rather than upgrading entire workspaces. API developers with latency-sensitive GPT-5.6 Sol workloads should evaluate Ultrafast.
Gemini / Google
★ Made by Google 2026 — Pixel 11 intelligence and Rambler speech cleaning
At the Made by Google event, Google introduced a suite of native Gemini features for Android and Pixel devices. Chief among these is Rambler, an AI-powered voice transcription tool that cleans up spoken language in real time by filtering out filler words, removing redundancy, and restructuring voice inputs into polished text before sending.
Google also rebranded and expanded its agentic assistant capabilities to Proactive Assistance (formerly Magic Cue), allowing Gemini to execute multi-step actions across more than 40 mobile apps — surfacing store loyalty cards based on precise location, parsing travel documents, and coordinating schedule additions in Calendar from Gmail. Gboard gained camera-based sign language translation.
Features: Rambler speech cleanser, Proactive Assistance, Gboard sign language translation | Processing: on-device offline dictation fallback; cloud-backed context processing for complex edits | App ecosystem depth: 40+ cross-app integrations | Accessibility: camera-based sign-language-to-text via Gboard | Availability: Pixel 11 series, August 20, 2026
Best for: Mobile users and professionals who rely on hands-free voice notes, automated daily admin tasks, and cross-app automation.
★ Gemini Omni Flash — multimodal video generation
Google rolled out Gemini Omni Flash, a high-speed multimodal generation tool focused on conversational video creation and editing. Building on Google's generative video research, Omni Flash allows creators and developers to generate, edit, and modify video clips directly using prompt-based controls with minimal rendering latency.
Model: Gemini Omni Flash | Modality: native text-to-video and interleaved multimodal editing | Deployment: Developer API and Gemini Advanced web | Latency profile: Flash-tier optimised for rapid generation cycles
Best for: Digital content creators, marketing teams, and developers building interactive video generation into creative workflows.
★ Gemini passes one billion users, expands to 14 third-party integrations
Google officially announced that the Gemini consumer application surpassed one billion users worldwide. Growth during the early weeks of August was significantly accelerated by large-scale telecom partnerships, including Reliance Jio in India.
User scale: >1,000,000,000 active users | Expansion: 14 enterprise and SaaS integrations across design, project management and hardware firmware | Hardware partners: XREAL (Android XR), Insta360 (Kira assistant)
Best for: Organisations looking for widespread user familiarity and integration across both mobile hardware and popular SaaS productivity platforms.
Microsoft Copilot
★ Copilot Ultrafast API and Premium Business seats
Microsoft previewed Ultrafast, a new API tier for GPT-5.6 Sol inside Copilot delivering up to 14× faster inference and 750 tokens per second on Cerebras infrastructure, retaining full GPT-5.6 Sol reasoning power. Separately, Copilot Business gained a Premium seat option.
Ultrafast: up to 14× faster, 750 tokens/sec, Cerebras infrastructure, same GPT-5.6 Sol capability | Premium seats: $125/user/month monthly, $100/user/month annual | Standard: $25 monthly, $20 annually | Premium capacity: 5× Standard, five-hour daily limit removed | Standard and Premium seats can be mixed in one organisation
Best for: Developers who previously had to compromise on model size for responsiveness, and teams with a small group of power users whose workloads justify higher capacity.
★ Copilot cybersecurity, Teens, Projects and Drive integration
Microsoft mirrored much of the OpenAI feature set across Copilot in the same window: GPT-5.6-Cyber for defensive cybersecurity with Daybreak Blue and Red access levels (available via Amazon Bedrock from August 11), a dedicated Copilot for Teens experience, interactive quizzes and project-only memory in Copilot Projects, Google Drive browsing in Copilot Library, and Computer History on macOS.
GPT-5.6-Cyber: Daybreak Blue (vulnerability discovery, secure-code review, malware analysis, incident response) and Daybreak Red (specialised, tightly governed) | AWS: both tiers via Amazon Bedrock from August 11 | Teens: ages 13–17 or system-estimated under 18; Study Mode, quizzes, Study Hours, responsible-homework reminders; no romantic language, no emotional dependence, no claims of consciousness; parent Quiet Hours | Projects: default ↔ project-only memory; shared Projects remain project-only | Drive: Docs, Sheets, Slides open alongside conversations, no manual re-upload | Computer History: admin-controlled, user opt-in, interaction events only, not available in EEA, UK or Switzerland | Restaurant reservations: OpenTable (global), Resy (US), Yelp (US/Canada)
Best for: Security teams needing frontier defensive capabilities under governance; families and educators; and Google Workspace users managing document-heavy collaboration.
⚠ Copilot — enterprise sync transition, Astra risk, upcoming retirements
Microsoft retired individually authorised sync connections for enterprise apps between August 10 and 14, moving organisations to admin-managed plugins. Internal evaluations suggesting Astra may reach Critical cybersecurity capability led to a pause in frontier reinforcement-learning runs, and two retirement dates are approaching.
Enterprise sync: individual connections retired August 10–14; organisations must now use admin-managed plugins for Google Drive, SharePoint and GitHub. GitLab Issues and Azure Boards replacements pending. Astra: August 7 disclosure, August 18 safeguards — frontier RL runs paused, monitoring expanded across Sol-level and Astra workloads, monitoring overhead estimated at 20% of inference compute. Retirements: OpenAI o3 leaves Copilot August 26 (API unaffected); DALL·E GPT retires August 30, replaced by Copilot Images. Ads expanded to the UK, Mexico, Brazil, Japan and South Korea on August 11.
Best for: Enterprise administrators ensuring continuity of connected knowledge sources, and users migrating workflows to GPT-5.6 and Copilot Images before the end of August.
Plans and pricing
Copilot Business Premium: $100/user/month annual, $125 monthly, 5× usage, no time limit | Standard: $20 annual, $25 monthly | Ultrafast: premium speed tier for GPT-5.6 Sol | Retirements: o3 (August 26), DALL·E GPT (August 30)
Best for: Identify the small number of employees whose usage justifies Premium rather than upgrading whole workspaces, and migrate remaining o3 and DALL·E GPT workflows before the end of the month.