OpenAI ships GPT-6.1 Sol for cheaper frontier coding and agentic work
OpenAI has started rolling out GPT-6.1 Sol in ChatGPT Work and Codex, with lower pricing and stronger agentic performance. The release points to a more practical path for scaling automated coding and work tasks.
OpenAI moves GPT-6.1 Sol into Work and Codex first
On September 29, 2026, OpenAI announced GPT-6.1 Sol and started rolling it out in ChatGPT Work and Codex. According to OpenAI, the new model improves on GPT-6 Sol in agentic coding, computer use, and professional work.
The rollout begins with Pro users and then expands to Plus, Business, Enterprise, and Edu. That matters because OpenAI is using its work-oriented surfaces as the first deployment path for the newer model, rather than treating it as a separate, niche release.
The main story is economics, not just capability
OpenAI’s DevDay 2026 materials say Sol is priced at one-fifth of Astra’s standard token prices. For teams that want to automate more coding and computer-use tasks, the shift is as important as the model upgrade itself.
OpenAI has been moving the GPT-6 line from flagship capability toward practical deployment economics. The newer Sol model fits that pattern by making frontier-style work more accessible for ongoing use, especially where many tasks need to run at scale.
Why developers will care about the tooling
According to OpenAI’s API materials, GPT-6 Sol is built for complex coding and agentic workflows. The API page also lists support for web search, file search, image generation, code interpreter, hosted shell, apply patch, Skills, computer use, MCP, and tool search.
OpenAI’s DevDay resource says Ultrafast is a paid speed tier with up to 8× faster token generation in Codex and 6× in the API. It also says the Agents API is in public beta with hosted execution, memory, tools, and multi-agent support. Taken together, those changes point to a platform that is meant to run more work continuously, not just answer prompts.
What changes for teams and admins
For organizations using ChatGPT Work, OpenAI says Enterprise and Edu workspace owners can control model access through Workspace settings > Permissions & roles. That gives admins a direct way to manage who can use the new model as it expands.
OpenAI also says Codex Cloud lets tasks continue from desktop, web, or mobile in reusable environments even while your computer is asleep. For teams, that suggests a workflow where coding and agent tasks can move across devices without forcing every step to stay tied to one active session.
What to watch next
OpenAI introduced GPT-6 Astra on September 3, 2026, with improvements in coding, research, computer use, and complex multi-step work. The newer GPT-6.1 Sol release builds on that direction while emphasizing lower-cost deployment and broader agent tooling.
The practical question now is how quickly teams can translate those capabilities into routine work. Readers should watch the rollout across ChatGPT Work, Codex, and the API, along with the expansion of the Agents API and the broader availability of Sol on each surface.
Sources
- ChatGPT — Release Notes | OpenAI Help Center OpenAI · September 29, 2026
- DevDay 2026 announcements and developer resources - Announcements - OpenAI Developer Community OpenAI Developer Community · September 29, 2026
- GPT-6 Sol Model | OpenAI API OpenAI · September 29, 2026
- Changelog | OpenAI API OpenAI · September 29, 2026
- The Work Now Within Reach | OpenAI OpenAI · September 8, 2026
This article was researched and drafted with AI from the sources above. Spot an error? Email info@hiddenproai.online.