OpenAI introduced GPT-6.1 Sol on September 29 for coding, computer use and professional work. Its standard API prices are $2 per million input tokens, $0.10 per million cached input tokens, and $10 per million output tokens.
The company describes performance near GPT-6 Astra on selected tasks at one fifth of Astra's standard input and output prices. Those are vendor results, not a promise of equivalent performance on every assignment.
A cheaper attempt still needs a useful result
Repeated agent work makes both failure rate and price matter. A model that needs fewer corrections may justify a higher price; a lower-priced model may make a bounded, repeatable task practical. Token pricing alone cannot settle that comparison.
Sol is available through the API and in Work and Codex on supported plans. The announcement says its Ultrafast mode is coming soon. Today's Astra Ultrafast launch should not be read as Sol Ultrafast already being available.
Included subscription allowances and metered API prices are different systems. These prices do not establish how many tasks a subscriber can run.