Loading article…
OpenAI unveiled ChatGPT Work powered by GPT‑5.6 and code suggests ads may appear in the free ChatGPT app, signaling new revenue and enterprise pushes.
OpenAI rolled out ChatGPT Work, an agentic mode that can automate multi‑step tasks across apps, while a beta Android build contains code references to an “ads feature,” indicating the company may soon monetize its free tier with advertisements【1】.
| At a glance | |
|---|---|
| Product | ChatGPT Work (agentic tool) |
| Model | GPT‑5.6 (latest frontier model) |
| Launch | Global rollout to Pro, Enterprise, Edu users (web & mobile) |
| Ad hint | Code in 1.2025.329 Android beta mentions “ads feature” |
OpenAI announced that ChatGPT Work can pull data from connected tools—Slack, Microsoft Teams, Google Drive, Salesforce, and more—then create finished outputs such as spreadsheets, slide decks, and dashboards without user prompting for each step【2】. The tool stays with a project for hours, breaking large goals into smaller tasks and asking for human approval before “sensitive actions”【3】. Powered by GPT‑5.6, the model family is marketed as delivering stronger reasoning across multi‑step coding, science, and cybersecurity tasks【2】.
The desktop app now bundles Chat, Codex, and Work into a single service available on Windows and macOS for all plans, including free users, while the standalone Codex app is folded into the updated desktop client【2】【3】. OpenAI also introduced “Sites” in public beta, letting users turn outputs into interactive web apps that update automatically when source data changes【2】【3】.
Separately, a discovered snippet of code in the 1.2025.329 beta of the Android app references “ads feature,” “search ad,” and “bazaar content,” suggesting OpenAI is preparing to display ads within the free tier that currently limits messages, memory, and reasoning capabilities【1】. The possibility aligns with earlier statements from CEO Sam Altman, who called ads a “last resort” but said the company is “not totally against them” and has been exploring ad integration based on user memory or chats【1】. No details were given on ad placement or pricing, and the beta version is not yet public.
ChatGPT Work targets enterprise customers where OpenAI has lagged behind rivals such as Microsoft’s Copilot and Google’s Gemini, which already embed AI agents into productivity suites. Analyst Holger Mueller notes that OpenAI must achieve rapid uptake and win user trust to match competitors’ traction【3】. The move to monetize the free tier with ads mirrors Google’s ad‑supported model and could provide a new revenue stream as OpenAI scales its user base, which now exceeds five million weekly Codex users and over one million for non‑software work【2】.
The dual announcements signal OpenAI’s push to broaden its product suite while seeking sustainable monetization. Whether ads will dilute the user experience or simply fund further AI development remains an open question as enterprise adoption accelerates.
Coverage is mostly measured — 179 of 201 reports stay neutral.
Every Monday — the token unlocks, Fed dates & catalysts set to move crypto and markets this week. So you’re never blindsided.
Free · 3-min read · one-click unsubscribe
AI-assisted synthesis by the TrendWatcher Editorial Desk · sourced from 3 outlets · Jul 15, 2026 · How we report
GPT-Red is an LLM designed to act as a super‑hacker for automated red‑team testing, helping OpenAI discover and patch vulnerabilities in its models before release.
When given the same task as human red‑teamers in a 2025 experiment, GPT-Red was more successful at finding effective attacks on an earlier GPT‑5 version.
OpenAI states that fewer than 23% of GPT-Red's strongest attacks succeed against GPT‑5.6, compared to over 90% success against GPT‑5.
OpenAI was founded on December 11, 2015 as a non‑profit to advance artificial general intelligence safely and beneficially, countering concerns that profit‑driven AI development could increase existential risks.
No, OpenAI has not released GPT-Red, citing concerns that the model is stronger than any potential copycat and could be misused.