OpenAI, the company behind the popular ChatGPT, offers a wide range of AI models, but the pricing structure can be complex for users. A recent analysis highlights the true cost of using ChatGPT, especially with its latest model, GPT-6 Astra. In Fast mode, a single request could cost up to $67, while in standard mode, it might reach $33. This is a significant jump compared to other models like Claude from Anthropic, which is often considered more cost-effective for similar tasks. Generative AI models operate using "tokens" rather than counting words. A token can be a full word, part of a word, punctuation, or other text fragments. Companies use tokens to set usage limits and charge based on consumption. The more tokens a model uses—typically due to longer responses or more complex reasoning—the higher the cost becomes, especially in pay-as-you-go plans. Previously, ChatGPT versions from 5 to 5.5 offered two main models: GPT-5.5 Instant for quick responses without reasoning, and GPT-5.5 Thinking, which included multiple levels of reasoning. With the release of GPT-5.6, OpenAI introduced three distinct model families: Sol, their flagship model; Terra, a balanced option for everyday tasks; and Luna, a faster and more affordable choice. Each can be configured with or without reasoning, similar to earlier versions. OpenAI also provides access to additional models through its API, including GPT-3.5-turbo, GPT-4, GPT-4-turbo, GPT-4o, and GPT-4.1. These models are typically used without reasoning capabilities. The recent launch of GPT-6 Astra, which includes reasoning like Anthropic's Fable, expands the available options to 14 models and 51 combinations depending on the reasoning level. This variety allows users to test different models based on their specific needs and budget. To compare performance and cost, the same prompts used for testing Claude were applied to ChatGPT. These included a riddle, a travel journal request, and a Chrome/Firefox extension idea. On average, ChatGPT prompts cost between $0.01 and $0.34 per request, depending on the model and complexity. This range reflects the variability in pricing based on the model's capabilities and the token usage required for each task.