I want to pay for a completed task, and a token discount alone cannot tell me what that costs. A cheaper call can still fund an endless meeting between agents. The calculation needs to reach a deliverable somebody can actually accept.
On August 21, 2026, OpenAI temporarily reduced GPT-5.6 Sol's price to $4 per million input tokens and $20 per million output tokens. Two days earlier, I had asked for parallel agents, each responsible for a separate part of the work. The discount speaks directly to that desire to distribute execution.
The trouble starts with how the work gets divided. If each agent must discover the same information before proceeding, part of the budget buys repetition. Then somebody has to compare the answers and resolve disagreements. That can justify its cost in a difficult investigation. For a simple assignment, coordination can end up outweighing useful work. Parallel execution needs a reason beyond an available slot.
In June, I had already suspected there were too many agents and suggested reducing the number. That moment of looking at the execution and questioning the excess remains a useful brake on enthusiasm. I want available capacity used, but watching every agent continuously consumes the attention I expected delegation to free.
Later, on September 3, 2026, I asked whether simple tasks were going to the cheapest model. That later question helps me revisit the August promotion. An economical choice starts with the result required and the cost of repairing a poor attempt. A cheap model that leaves work unfinished may need another run with additional context.
I prefer splitting activities that return something independently useful. A search can bring back evidence while another part advances. Two agents renegotiating the same decision at every step need a different arrangement. The number of executors becomes interesting only after each responsibility is clear enough to explain.
To assess the promotion, I will compare the cost of reaching the same deliverable, including discarded attempts and the attention required to coordinate. Multiplying the new price by imagined consumption helps planning; personal savings require observed execution. I will keep the advertised price separate from that result and leave the selection rule open to revision. The promotion is temporary, and I do not want to inherit an expensive configuration just because it looked cheap in the month of the announcement.