What changed
On October 7, 2026, OpenAI cut its GPT-6 Sol API price from $4 to $2 per million input tokens and from $20 to $10 per million output tokens. GPT-6 Luna, the model heading to free users, dropped to $0.10 input and $0.50 output per million tokens. This is a 50% cut on the flagship paid model and roughly 58% on the output side for the free-tier model.
To put that in context: generating 100,000 words of customer-support replies now costs about $1.50 instead of $3. Processing 10,000 customer emails for sentiment tagging costs about $0.30 instead of $0.60.
What you can do with it now
If a task felt too expensive to automate last month, run the numbers again. Customer replies: draft responses to common questions in bulk. Quote generation: feed product specs and customer requirements into a prompt, get a first draft back in seconds. Data extraction: pull invoice details, contact info or line items from PDFs or scans without hiring a VA.
You can connect to the OpenAI API through tools like Zapier, Make, or a developer using a few lines of Python. Most workflows cost under $10 a month at this price if you're processing a few hundred documents or messages a week.
What to keep in your own hands
Don't paste in customer payment details, employee contracts, or anything with a social-security number. Use the API for draft work, not final decisions—a person reviews the quote before it goes out, checks the extracted data before you file it, and reads the reply before you hit send. And run a test month: set a $50 budget cap in your OpenAI account so a runaway script can't rack up a surprise bill.