OpenAI shipped GPT-6.1 Sol today, a lower-cost model the company says comes close to matching its flagship GPT-6 Astra.
The company says Sol brings real gains over the prior version on complex, professional-grade tasks: writing and debugging code, parsing dense documents, and carrying out multi-step business workflows. OpenAI's framing is specific - this isn't a chatbot upgrade, it's aimed at people who use the model as a tool inside longer, higher-stakes work. The headline pitch is near-parity with Astra, OpenAI's top-tier model, at a lower price.
That price angle is the real story here. Every major lab is under pressure to prove its models are worth what they cost, and "almost as good, for less" is becoming the standard move once a flagship has been out long enough to need a budget sibling. It also puts pressure on rivals selling similarly priced mid-tier models on capability alone.
"Nearly matches" is OpenAI's own claim, not an independently verified one - no third-party benchmark numbers came with the announcement. Treat it the way you'd treat any vendor's self-graded homework, until someone outside OpenAI runs the numbers.