Model naming

Every model has an ID. An optional variant changes how much context the model sees and how it is billed. Browse them all in the Models section.

claude-sonnet-5 model -cheap extra :metered variant

Variants

Variant	Context	Billing
:free	Capped at 5,000 tokens	1 daily free credit per request
premium	Set by your chat memory settings	The model's price for the first 15,000 tokens, then one more per 5,000
:metered	Set by your chat memory settings	The model's price per 5,000 tokens, from the first token
:full-context	The model's full native window	One flat price per request, whatever the length
:metered:full-context	The model's full native window	The model's price per 5,000 tokens, from the first token
:32k-context, :64k-context, :128k-context, :256k-context	An approximate input budget of 32,000, 64,000, 128,000, or 256,000 tokens	One flat price per request; if combined with `:metered`, the model's price per 5,000 tokens instead
:payg	The model's full native window	Per token: the model's price per 1M input, output, and cached tokens (see Credits)
:free with a context suffix	The context variant's budget, subject to the model's native limit	1 daily free credit per request

Context variants use their own budget instead of your plan's chat memory sliders. Named budgets are approximate, and the model's native limit still applies. Copy a complete ID from Models; adding a suffix does not create an available variant.

Extras

Name	Meaning
`-cheap`	Cheaper price
`-code`	Tuned for coding
`-thinking`	Forces thinking
`-non-reasoning`	Turns thinking off
`-flatcost`	One flat price
`-official`	Runs on the provider's own API

To see what a request actually costs, use the cost calculator. The three credit types a request can spend are explained on the Credits page.