OpenAI introduces cheaper GPT-6.1 Sol model following shelved Astra plans

UPDATE - follow-up to OpenAI unveils GPT-6.1 Sol and 'Dots' AI agents at DevDay: OpenAI launched GPT-6.1 Sol at its San Francisco DevDay conference on Tuesday, September 29, 2026. Sam Altman presented the model as a lower-cost option that provides near-Astra intelligence at one-fifth of the price for coding and building tasks. He noted that Sol is smarter than Astra in certain ways and will serve as a daily workhorse that helps agents reuse context and handle longer tasks cheaply.

Pricing and Features

Developer documentation shows that GPT-6.1 Sol costs $2 per million input tokens and $10 per million output tokens. It supports reasoning levels from low to maximum, features a 1.05-million-token context window, and can generate up to 128,000 tokens per response. The model supports tools like web search, file search, code interpretation, hosted shell, computer use, image generation, skills, and Model Context Protocol. Users can access it through the Responses API, which also supports function calling and structured outputs.

Enterprise Pilots and Safety Shift

OpenAI has started enterprise pilots for specialist Dots, which possess their own identities, credentials, and system access to handle defined jobs. The company is collaborating with Microsoft to build governance and security controls into Agent 365 for these agents. These announcements followed OpenAI's confirmation that it shelved GPT-6.1 Astra, previously slated for October, because internal tests revealed safety and alignment issues regarding authorized scope and communication with users.

Leadership Vision

During his closing remarks, Mr. Altman rejected the comparison of AI to an industrial revolution and said he finds the idea of making people cogs in a machine off-putting. He stated that parts of life should not be automated and argued that proper AI development could instead bring a new renaissance that gives people more tools to create and learn.

Context

Artificial intelligence companies regularly adjust their product roadmaps and pricing models to balance performance and operational expenses for developers. Introducing lower-cost models helps make advanced capabilities accessible to a wider range of enterprise and individual users. Such adjustments reflect the ongoing challenges of meeting safety standards while expanding commercial offerings in the tech industry.