A cost analysis of GPT-5.5 reveals users are paying 49-92% more than with GPT-5.4, despite the new model generating shorter responses for certain prompts. OpenAI launched GPT-5.5 with a 2x price increase—input tokens jumped from $2.50 to $5.00 per million, output tokens from $15 to $30 per million.

Researchers used OpenRouter data to track users who switched from GPT-5.4 to GPT-5.5, comparing the same workflows across both model versions. The analysis found GPT-5.5 produces 19-34% fewer completion tokens for prompts over 10K tokens, partially offsetting the price hike. However, shorter prompts under 10K tokens see minimal reduction in output length while still incurring the full price increase.

Cost impact varies significantly by use case. Users with prompts under 2K tokens face the steepest increases at 92%, while those with 50K-128K token prompts see the smallest increase at 49%. Organizations processing large documents benefit from the model's conciseness, but lightweight use cases bear the brunt of the pricing shift.