Knowledgebase

You choose a model by its alias. Start with default and move up only when the answers fall short.

• default: the fast everyday model and the cheapest. Good for most questions, scripts and coding assistants.
• large: a larger, stronger model for harder tasks. It costs several times more per token.
• alt and alt-large: a second pair of models with a different style. alt is very cheap on repeated input, which suits long conversations and agents that resend the same text.
• xl: for the hardest reasoning and coding tasks, when large and alt-large fall short. It is the most expensive model for input and output, and it always reasons before it answers.
• The nolimits- models: fewer refusals. nolimits-story is tuned for fiction and creative writing.

Every form of default, large, alt and alt-large, the -instant and -think ones included, has a second provider behind it for chat and Gemini-format requests: if the first one fails before the answer starts, the second takes the request at the same price. xl and the nolimits- models have one provider each.

-instant and -think

Some aliases come in two extra forms. -instant answers straight away. -think reasons at length before answering, which helps on hard problems but costs more, because reasoning counts as output. xl always reasons, so on xl-instant the reasoning is kept short rather than skipped.

How long can a conversation be?

Most models take about a million tokens of context. nolimits-story takes 128K tokens.

The full list

Every alias, the model behind it, its prices and its context length: Models and what a request costs. Each model can also be called by its own name at the same price. Your key can list them all, with current prices:

curl https://virsilipas.mcx.fi/v1/models -H "Authorization: Bearer $KEY"