Part 3/9:
Expectations around cost-effective usage were dashed when examining the pricing model for GPT-4.5. Contrary to assumptions that it would be cheaper, the input costs $75 per million tokens and outputs cost $150, making it one of the most expensive options on the market. In comparison, reasoning models previously available have significantly lower costs, emphasizing this model's positioning as a premium but not necessarily superior choice.
Moving beyond mere functionality, the model’s high cost raises fundamental questions about its economic viability for developers and businesses, particularly when compared to other more efficient models.