Build more. Spend less.
A great model for each job. A price that makes sense.
All tracked families
Intelligence
Providers
No measured, priced models in this view.
75% input / 25% output
Your two-model toolkit
One big thinker. One everyday builder.
Or the same model for
both.
Data & unmeasured models
Planning uses general intelligence. Coding uses SciCode results. These are different measurements, not comparable scores.
Prices can be shown per 1M tokens (blended input/output rate) or per task (the measured cost to complete one Intelligence Index task). The two bases are different measurements; compare models within one basis only. A per-task cost is not derived from your token mix and is published for fewer models.
| Model | Why it is not plotted | Selection |
|---|