OpenAI and Anthropic news from the past week (Week 20, 2026)

OpenAI launched the Deployment Company, Daybreak, Codex in the ChatGPT mobile app, and personal finance in ChatGPT, Anthropic launched Claude for Small Business, Claude Platform on AWS, Claude for the legal industry, and reportedly in talks to buy Stainless, and more

OpenAI launched the OpenAI Deployment Company with Tomoro acquisition, shared Daybreak vision for cybersecurity, brought Codex into the ChatGPT mobile app with Remote SSH, rolled out a preview of personal finance in ChatGPT to Pro users in the US, added the OpenAI Developers plugin in Codex, expanded file library and storage management to Free and Go users, rolled out more inline images from the web in ChatGPT answers for Free users, shared updates on helping ChatGPT better recognize context in sensitive conversations, opened a Campus Network student club interest form, offered two free months of Codex for new enterprise users, partnered with Malta to bring ChatGPT Plus to all citizens, shared an update on their response to the TanStack npm supply chain attack with macOS certificate rotation, shipped Codex desktop app updates with customizable shortcuts and performance gains, published engineering posts on the Codex Windows sandbox and Parameter Golf takeaways, deprecated DALL-E model snapshots and the Realtime API Beta, reorganized leadership with Greg Brockman now leading product strategy, plus a simplified “Intelligence Picker” model picker was spotted, ChatGPT memory replacement plans, reports that OpenAI is considering legal action against Apple over their ChatGPT partnership, and reports that OpenAI quietly acquired voice cloning startup Weights.gg

Anthropic launched Claude for Small Business with PayPal partnership, made Claude Platform on AWS generally available, released Claude for the legal industry with 20+ MCP connectors and 12 practice-area plugins, shipped agent view, fast mode for Opus 4.7, and the /goal command in Claude Code, introduced a dedicated monthly Agent SDK credit for paid plans starting June 15, raised Claude Code weekly limits by 50% through July 13, formed a $200M partnership with the Gates Foundation, released Claude’s Constitution as an audiobook, published a cybersecurity case study on the CLUE threat detection platform, shared a paper on US-China AI competition scenarios for 2028, and is reportedly in advanced talks to acquire developer tools startup Stainless

OpenAI launched the OpenAI Deployment Company with $4B initial investment, led by TPG with 19 partners, to embed Forward Deployed Engineers into organizations, and agreed to acquire Tomoro to bring ~150 FDEs from day one

[embed]

Anthropic released Claude’s Constitution as an audiobook, read by two of its authors Amanda Askell and Joe Carlsmith, with a Q&A on the writing process and how the document might change

[embed]

OpenAI shared Daybreak, their vision for cybersecurity built around Codex Security, with a new GPT-5.5 Trusted Access for Cyber tier alongside GPT-5.5-Cyber for authorized red teaming and penetration testing

[embed]

Anthropic introduced agent view in Claude Code as a research preview, a single place to manage parallel sessions, background tasks with /bg, and steer many agents at once

[embed]

Spotted OpenAI testing a simplified “Intelligence Picker” in the ChatGPT web prompt composer that replaces separate model and reasoning effort selection with Instant, Medium, High, and Pro options

[embed]

Anthropic made the Claude Platform on AWS generally available, bringing the full set of native Claude API features with AWS IAM authentication, CloudTrail logging, and single AWS invoice billing

[embed]

OpenAI added the OpenAI Developers plugin to Codex so it can help build AI apps and agents faster using OpenAI APIs

[embed]

OpenAI opened a Campus Network student club interest form to partner with student clubs across universities on hands-on AI learning, events, early access, and student leader connections

Anthropic released Claude for the legal industry with 20+ new MCP connectors across contract lifecycle, document management, e-discovery, and legal research, plus 12 practice-area plugins, and partnerships with the Free Law Project and Justice Technology Association

OpenAI expanded inline images from the web in ChatGPT answers to Free users globally on web and iOS for responses using GPT-5.5 Instant, including well-known people, places, and products

Anthropic launched fast mode for Claude Opus 4.7 in research preview on the API and in Claude Code, with 2.5x faster responses at $30/$150 per MTok, becoming the default fast mode model on May 14

[embed]

OpenAI shared what Parameter Golf taught them, an open ML challenge with 1,000+ participants and 2,000+ submissions, highlighting top record and non-record track submissions and the widespread use of coding agents

[embed]

The Information reported Anthropic is in advanced talks to acquire developer tools startup Stainless for at least $300M, which builds SDK generation software used by OpenAI, Google, and Anthropic

[embed]

Anthropic published a case study on how their cybersecurity team built CLUE (Claude Looks Up Evidence), a detection and response platform powered by Claude that automates alert triage and investigations

OpenAI deprecated DALL-E model snapshots (dall-e-2 and dall-e-3) and the Realtime API Beta on May 12, recommending gpt-image-2, gpt-image-1, or gpt-image-1-mini and the released Realtime API instead

OpenAI shared an update on their response to the TanStack npm supply chain attack (Mini Shai-Hulud), confirming no user data or production systems were compromised but rotating macOS code-signing certificates by June 12

Anthropic launched Claude for Small Business, a toggle install in Claude Cowork with connectors for QuickBooks, PayPal, HubSpot, Canva, Docusign, Google Workspace, and Microsoft 365, plus 15 ready-to-run workflows, a free AI Fluency course with PayPal, and a 10-city SMB tour

Anthropic added the /goal command to Claude Code, matching the Codex /goal feature, which sets a completion condition and keeps Claude working across turns until a small fast model confirms the condition is met

[embed]

Anthropic announced that starting June 15, paid Claude plans get a dedicated monthly Agent SDK credit for programmatic usage (Pro $20, Max 5x $100, Max 20x $200, Team Standard $20/seat, Team Premium $100/seat), with subscription limits reserved for interactive use

[embed]

OpenAI is offering eligible enterprise customers two months of free Codex usage for new users if they switch in the next 30 days

[embed]

Anthropic increased Claude Code weekly limits by 50% through July 13, applied automatically across CLI, IDE extensions, desktop, and web, stacking with last week’s 2x increase to 5-hour limits

[embed]

OpenAI published an engineering deep-dive on building a safe, effective sandbox to enable Codex on Windows, walking through the unelevated and elevated sandbox designs that combine SIDs, write-restricted tokens, firewall rules, and custom users

[embed]

Anthropic announced a $200M partnership with the Gates Foundation over four years across global health, life sciences, education, and economic mobility, including work on polio, HPV, and preeclampsia, K-12 tutoring, and smallholder farming

[embed]

OpenAI expanded file library and storage management in ChatGPT to Free and Go users including the EEA, with storage limits ranging from 500 MB on Free to 100 GB on Pro

OpenAI shared updates on helping ChatGPT better recognize context in sensitive conversations using safety summaries, with safe-response performance improving 50% in suicide and self-harm cases and 16% in harm-to-others cases in long single-conversation scenarios

[embed]

Anthropic published a new policy paper outlining two scenarios for US-China AI leadership in 2028, calling for tightened export controls, disrupting distillation attacks, and accelerating democracies’ AI adoption to lock in a 12–24 month lead

[embed]

Spotted OpenAI planning to replace saved memories in ChatGPT with a structured summary on September 1, plus new mentions of ChatGPT in Viber, sending email directly from ChatGPT with a connected provider, and Finances in ChatGPT rolling out to Pro users in the US

[embed]

OpenAI brought Codex to the ChatGPT mobile app in preview on iOS and Android, letting you connect to Codex running on your Mac mini, Mac Studio, or MacBook to monitor and continue tasks away from your keyboard, also making Remote SSH and Hooks generally available

[embed]

The Information reported OpenAI is considering legal action against Apple over breach of contract related to the ChatGPT integration deal announced two years ago, citing Apple’s growing closeness with Google and Anthropic

[embed]

OpenAI rolled out a preview of a new personal finance experience in ChatGPT to Pro users in the US on web and iOS, with Plaid-powered account connections, a finances dashboard, and GPT-5.5 Thinking as the default model

[embed]

WIRED reported OpenAI is reorganizing with Greg Brockman officially leading product strategy alongside infrastructure, merging ChatGPT and Codex into one core product team, with Thibault Sottiaux leading core product and platform, and Nick Turley moving to enterprise products

[embed]

OpenAI shipped Codex desktop app updates with customizable keyboard shortcuts, Git actions moved back into the review flow, an improved thread panel, a cleaned up local server list, and performance gains including ~75% less re-rendering when switching threads

[embed]

The New York Times reported OpenAI quietly acquired Weights.gg earlier this year, a startup that offered AI tools to create voice clones, with employees disbanded across the company and no plans to release a similar product

[embed]

OpenAI partnered with the Government of Malta to offer ChatGPT Plus free for one year to all Maltese citizens who complete an AI literacy course developed by the University of Malta

That’s all for this past week, catch you in the next recap. And if you spend a lot of time in ChatGPT and Claude, you’ll get more out of both with AIPRM, our browser extension with thousands of prompt templates, writing styles, tones, and teams collaboration, at AIPRM.com.

Follow me on X/LinkedIn/Threads/Bluesky for more