OpenAI has announced GPT 6.1 Sol, a new model it describes as offering near-Astra intelligence at a fifth of the price. The release was published on OpenAI's website and quickly picked up by AI commentator Simon Willison, who covered it in a blog post dated 29 September 2026.
The framing is simple and striking: much of the performance of Astra, OpenAI's top-tier intelligence, for roughly 20% of the cost. That combination of capability and price is what makes GPT 6.1 Sol a notable moment for anyone building with or buying AI.
GPT 6.1 Sol pricing and positioning
The headline claim from OpenAI is straightforward: GPT 6.1 Sol delivers near-Astra intelligence at a fifth of the price. In practice, that positions the model as a cheaper route to high-end AI performance, rather than a new top-end tier above Astra.
Both OpenAI's own announcement and Simon Willison's write-up use the same description, so there is no disagreement between the sources on the core claim. What has not been detailed in the available material is the exact price figures or the specific benchmark scores behind the comparison.
What testers and observers are saying
Simon Willison's post notes he was late with his usual "pelicans" test because he was live-blogging OpenAI's DevDay 2026 keynote. He published the pelican results for GPT-6.1-Sol alongside his commentary.
According to Willison, the pelicans generated by GPT 6.1 Sol are not notably different from those of the wider GPT-6 family. That is an informal observation rather than a formal benchmark, and it is the only hands-on detail available from the sources so far.
What it means for you
For developers and businesses in Tunisia, the main takeaway is cost. If GPT 6.1 Sol really does offer near-Astra intelligence at a fifth of the price, it could lower the budget needed for advanced AI features in apps, customer service tools and internal automation.
