OpenAI restructures pricing and launches Dots agents on its path to a 2027 public offering
OpenAI opened the developer conference “Dev Day 2026” at the Fort Mason Center in San Francisco, attended by 2,500 participants, unveiling more than twenty technological launches that define its infrastructure roadmap and commercial direction. The most prominent announcement was the Dots agents, software agents that continue working and executing operations after a user closes the ChatGPT interface. This shift reflects a move in interaction engineering from real-time chat sessions to independent runtime environments that accompany enterprise workflows without interruption, redefining the role of intelligent models within daily work environments.
Regarding models and software execution, the company announced pricing for the GPT-6.1 Sol model via the API at about $2 per million input tokens and $10 per million output tokens, with the cost of temporarily cached inputs reduced to just $0.10 per million tokens. In a parallel move, the company updated its professional subscription plans to three monthly tiers priced at $100, $200 and $500. At the same time, OpenAI confirmed that the launch of the GPT-6.1 Astra model is on hold because of safety and security concerns, while unveiling a Private Intelligence system aimed at large enterprises that gives institutions strict control over their sensitive data when running the most advanced models in collaboration with major clients.
These operational decisions coincide with an increasingly clear financial trajectory for the company; Sarah Frair, OpenAI’s chief financial officer, told employees that the company is focused on becoming a publicly listed firm in 2027, or sooner if the business curve continues its rapid ascent. In the context of eager financial markets, the Kraken platform began offering perpetual, cash-settled futures contracts linked to the valuations of OpenAI and Anthropic ahead of the IPO, with leverage up to five times, while completely excluding traders in the United States, the European Economic Area, Canada, Australia and New Zealand. At that time, the company denied any intention to issue its own digital currency, reaffirming its explicit condemnation of fraudulent products that previously exploited its brand.
For those who build and manage tech solutions in the Gulf, Egypt or the broader Arab region, this landscape imposes direct adjustments to cost calculations and software design. Reducing the cost of temporarily cached inputs to $0.10 per million tokens shrinks the inference bill for banks and regional service firms that rely on long contexts and repetitive data, making back-end automation more economically viable. Likewise, the emergence of permanent agents such as Dots shifts regional developers’ skills from crafting instantaneous commands to building monitoring controls and managing access permissions for independent systems, while a private data system for government and financial institutions subject to local compliance rules provides a path to test advanced models without sacrificing corporate confidentiality.
On the investment side, limiting the trading of derivative contracts that value the company to non-Western markets gives regional traders and investment offices a high-risk speculative tool that anticipates the 2027 offering without traditional regulatory oversight.The shelving of the Astra model and the offering of packages up to $500 per month send a decisive message: tool efficiency is now measured by immunity, compliance and the management of costs for agents that operate continuously.