Llm pricing is a pricing and specification comparison platform for large language models (LLMs). Its main purpose is to help users compare the input, output, and total token costs of GPT, Claude, Gemini, Llama, and other LLMs, and to estimate API costs based on usage.
In terms of main capabilities, the platform provides a list of available models, supports filtering by provider, and allows sorting by cost columns. Users can enter expected input and output token usage: where M means one million tokens, and an input of 1 represents 1,000,000 input or output tokens, and the cost estimate updates instantly. The estimation logic is: the number of input tokens multiplied by the model's input rate, the number of output tokens multiplied by the output rate, and then the two are added together to get the total amount.
In terms of usage, users can set expected input and output usage, filter by provider, and sort by cost columns; at the same time, they can compare context length, supported modalities, and task quality. The platform notes that the lowest token rate is not always the lowest total cost.
The applicable scenarios are mainly API cost estimation and model selection reference, for example, comparing the rates of different models and providers based on expected workload.
In terms of limitations, the platform states: provider-specific fees, discounts, caching, tools, images, and other add-ons may be billed separately; providers usually price generated output higher because generation requires sequential computation; rates vary by model and provider; model specifications and prices come from an upstream model catalog, and when the source service is available, they are refreshed on a six-hour cache cycle; rates may change, and the provider's current pricing should be verified before purchase.
More details about the billing model, pricing, platform form, and so on are not yet provided in the materials.
