Models

Every model, priced up front.

Klara Base Preview v1 and v2 are smaller early versions of the new, more powerful Klara models we are still finishing. They are the meantime releases; price, access, context, and live status are shown before a request can run.

Klara Base Preview v1

StandardTemporarily unavailable

A smaller preview of the new, more powerful Klara models now being completed. Preview v1 prioritizes directness and speed.

StandardLong contextSecurity researchBio literatureDirectAgenticTools

Best for: Direct, fast research and technical work while the next generation of Klara models is being completed.

Input: $2.50/M
Output: $8.00/M
Request minimum: Set at launch
Context: 203k
Access: Standard
Retention: 30-day ops
Route: AdvancedMind AI managed routeUnavailable

Klara Base Preview v2

StandardTemporarily unavailable

A smaller preview of the new, more powerful Klara models now being completed. Preview v2 prioritizes capability retention on harder work.

StandardLong contextSecurity researchBio literatureDirectAgenticTools

Best for: Harder reasoning, coding, and research work with capability retention prioritized.

Input: $2.50/M
Output: $8.00/M
Request minimum: Set at launch
Context: 203k
Access: Standard
Retention: 30-day ops
Route: AdvancedMind AI managed routeUnavailable

A cost and route record on every response.

response recordper response

request_id: unique per response
model · route: id + AMAI route served
tokens: input / output · metered
access tier: standard / premium research
retention: per-model mode
cost: prepaid debit

Pricing is shown before you run a model.

Reserved Preview GPU capacity uses purchased prepaid credits. Billing enables only a signed-readiness-qualified top-up; membership checkout is paused and plan allowance does not fund a Preview reservation. Apply for Research Mode only if the work needs an approved higher-risk project scope.

Check Preview funding How Research Mode works