- OpenAI released ChatGPT Work on July 9, a GPT-5.6 product that turns notes, drafts, and reference files into finished documents, spreadsheets, and decks.
- The company simultaneously retired its standalone Atlas browser, with a shutdown date of August 9, and folded its capabilities plus the Codex desktop app into a single unified ChatGPT application.
- GPT-5.6 is priced across three tiers, from Luna at $1 input and $6 output per million tokens to the Sol flagship at $5 and $30, and Sam Altman says it is 54 percent more token efficient on agentic coding.
OpenAI collapses its tools into one surface built to finish work
OpenAI released ChatGPT Work on July 9, a GPT-5.6 product that takes unstructured input and returns a completed deliverable rather than a chat reply that still needs assembling. A user drops in meeting notes, half-written passages, and reference files, and the product returns a formatted document, a working spreadsheet, or a presentation draft that follows the user's own templates.
The launch came with a consolidation. According to OpenAI, the company is discontinuing its standalone Atlas browser less than a year after launch, with a formal shutdown on August 9, and is merging its capabilities and the separate Codex desktop app into a single unified ChatGPT application. The prior desktop app was renamed ChatGPT Classic.
All these capabilities were built on what we learned from Atlas users who took a leap of faith on a new browser. You taught us how agents can help make browsing and doing work on the open web better, and we're bringing those learnings into ChatGPT.James Sun, OpenAI
The pricing tells its own story about where OpenAI expects the volume to come from.
| Tier | Input / 1M tokens | Output / 1M tokens | Role |
|---|---|---|---|
| Sol | $5 | $30 | Flagship |
| Terra | $2.50 | $15 | Cost-competitive with GPT-5.5 |
| Luna | $1 | $6 | Fastest and cheapest |
Why folding everything into ChatGPT puts OpenAI in Microsoft's lane
The strategic message is that OpenAI no longer wants ChatGPT to sit beside your work. It wants ChatGPT to be the place the work gets produced. That is the same claim Microsoft makes for Copilot inside Word, Excel, and PowerPoint, and Copilot runs on OpenAI's models. Within days of the launch, Google framed its Gemini Enterprise platform as a direct response, a sign the contested ground has moved from model benchmarks to the output layer where documents actually get made.
Every enterprise now is thinking about spend and the value they're getting in exchange for AI.Sam Altman, OpenAI CEO, to CNBC
A single app that produces finished output, priced from a dollar per million tokens at the low end, is built for exactly that scrutiny. Retiring Atlas removes a product that fragmented attention and rolls its agentic browsing into the surface OpenAI wants people to live in. The consolidation is also a cost calculation, aimed at giving enterprises one place to measure what they spend against what they get.
ChatGPT Work is not a better chatbot. It is OpenAI's move to control where output is produced, and by killing its own browser to get there, the company is signaling that it will cannibalize its own products before it lets the work happen anywhere else.
Santage is committed to independent, transparent journalism. This article is produced in accordance with Santage's Editorial Standards and aims to provide accurate and timely information. Readers are encouraged to verify information independently.