OpenAI has introduced ChatGPT Work, an intelligent workplace agent paired with three new GPT-5.6 models, signaling a shift from simple chatbots to AI systems that can handle complex workflows, draft documents, and manage projects autonomously. The launch, announced on July 9, 2026, includes Sol (the flagship model), Terra (a balanced mid-tier option), and Luna (a fast, cost-efficient variant). Meanwhile, a real-world experiment shows how ChatGPT is already changing how people consume information and manage their time.

What Makes ChatGPT Work Different From Regular ChatGPT?

ChatGPT Work is not just another feature update. It represents OpenAI's vision of moving beyond simple "chat" interfaces toward agentic systems that take initiative, maintain context across long workflows, and produce finished deliverables rather than raw text. The agent is designed to turn scattered notes, drafts, and ideas into polished work by integrating context from your team's existing tools.

One journalist at Tom's Guide tested this concept in an unconventional way. Instead of doomscrolling through social media for 48 hours, they used ChatGPT to catch up on news by asking it to generate a comprehensive summary prompt. The chatbot created a structured breakdown covering the biggest stories, AI and tech developments, entertainment, internet buzz, stories worth ignoring, winners and losers, and a week-ahead preview. The experiment proved surprisingly effective, helping the journalist stay informed while breaking a harmful habit.

How to Use ChatGPT as a Personal News Curator

• Create a Structured Prompt: Ask ChatGPT to summarize news from a specific time period using categories like "Biggest Stories," "AI and Tech," "Entertainment," and "Looking Ahead," prioritizing signal over noise rather than engagement-driven sensationalism.
• Request Source Links: Have ChatGPT include links to original sources and provide search terms you can use to conduct your own follow-up research on topics of interest.
• Filter for Relevance: Ask the chatbot to identify stories that are genuinely important versus those that generated attention but lack lasting significance, helping you focus on what actually matters.
• Set Time Boundaries: Use ChatGPT to catch up on news from a defined period (48 hours, one week) rather than endless scrolling, creating a healthier information consumption habit.

The journalist noted that ChatGPT "did a great job filtering its information rollout for high signal over high engagement," focusing on stories most likely to matter a month from now rather than what dominated social media feeds. This approach revealed important developments like U.S.-Iran tensions, Google's Pixel 11 launch plans, and even a warning about fake GTA 6 pre-order scams.

Understanding the Three New GPT-5.6 Models and Their Pricing

OpenAI's decision to release three distinct models under the GPT-5.6 umbrella reflects a maturation of its product strategy. Rather than forcing users into a one-size-fits-all approach, the company has embraced what it calls "intelligence stratification," offering different tiers of capability at different price points.

Sol is OpenAI's most capable model to date, representing a qualitative leap in reasoning depth and multi-domain mastery. The company reports that Sol is 54% more token-efficient in AI coding tasks compared to previous frontier models, meaning it accomplishes the same coding work using roughly half the computational tokens. This translates to faster results, lower costs, and reduced energy consumption. Sol is also positioned as OpenAI's "strongest cybersecurity model yet," with frontier-level performance in threat modeling, code review, blue teaming, and incident response coordination. Pricing for Sol is $5 per million input tokens and $30 per million output tokens.

Terra is the balanced, lower-cost option designed for everyday professional and personal use cases where Sol's full frontier capability may be unnecessary. Terra's performance is characterized as "competitive with GPT-5.5," which was, until this launch, one of the most capable AI systems in existence. For the vast majority of real-world tasks, Terra will feel indistinguishable from a frontier model. It costs $2.50 per million input tokens and $15 per million output tokens, making it accessible to students, small businesses, content creators, and developers working on standard coding tasks.

Luna is OpenAI's fastest and most cost-efficient model, priced at just $1 per million input tokens and $6 per million output tokens. Luna is designed for high-volume, latency-sensitive, cost-conscious deployments such as real-time chat applications, customer support automation, content moderation at scale, and rapid prototyping where iteration speed matters more than final polish.

Why This Matters for How You Work and Consume Information

The combination of ChatGPT Work and the three GPT-5.6 models signals a broader shift in how AI is becoming integrated into daily workflows. Rather than treating AI as a novelty tool for occasional use, OpenAI is positioning these systems as permanent workplace companions capable of handling complex, multi-step tasks. The real-world experiment with news curation demonstrates that this shift is already happening in practical ways.

The journalist's experience revealed three key insights: doomscrolling first thing in the morning is a significant detriment to mental well-being and productivity, using ChatGPT to build a quick summary of major news stories is incredibly helpful, and using the chatbot's report to conduct independent research is equally beneficial for both personal and professional purposes. This suggests that AI tools are not just replacing human effort but fundamentally changing how people approach information consumption and time management.

For enterprises, the cybersecurity capabilities of Sol could justify adoption costs alone. For individuals and small businesses, Terra and Luna make advanced AI accessible at price points that were previously impossible. The introduction of ChatGPT Work as an intelligent agent capable of drafting documents, creating spreadsheets, and building presentations suggests that the next phase of AI adoption will focus less on "chatting" with AI and more on delegating actual work to autonomous systems.