OpenAI shipped a near-top model at a fifth of the price, and held the top one back
At DevDay on 29th Sep'26, OpenAI released GPT-6.1 Sol, which it describes as near-Astra intelligence at one-fifth of Astra's standard token cost. It had halted the planned October launch of GPT-6.1 Astra the day before, after the model failed safety testing. GPT-6 Astra, released on 3rd Sep'26, remains the top model and powers the new Dots agents. The performance claim is OpenAI's own, and independent benchmarks were not available when this was written.
What it actually means
The price is the news for anyone watching a token bill: a fifth of the top-tier rate for a model OpenAI calls close to it. That is a claim to test before you believe it. Run your hardest real prompts through Sol and through whatever you use now, because near is doing a lot of work in that sentence and vendor benchmarks are chosen by the vendor.
The delay matters as much. The best model is no longer guaranteed to be the next one out, so a roadmap that assumes the flagship upgrade lands on schedule needs a fallback to the tier below it, which makes Sol the model most teams will actually build against this quarter.
The top model slipped; the one at a fifth of the price didn't.
1/5of Astra's token cost, OpenAI's claim for Sol
Source: The Next Web · Thursday 1 October 2026