Quick answer
Factory's tiers differ in how much usage they include and which models they unlock, not in how individual requests are measured. Standard is aimed at regular individual use and Core at heavier or organization work with higher allowances. Verify the current allowances on Factory's pricing page before committing, since tiers change.
What differs
| Dimension | Standard | Core |
|---|---|---|
| Audience | Individual developer | Heavy or team use |
| Allowance | Moderate rolling window | Larger window |
| Model access | Included set | Broader set where offered |
| Admin | Personal settings | Organization controls |
| Overflow | Wait or upgrade | Upgrade or additional capacity |
Decide with data
- Run one full billing period on the lower tier.
- Count the days you were blocked by the window.
- Note which models you actually used.
- If blocks cost more than the price gap, move up.
- If blocks are rare, stay and pace the work.
Counting blocked days is more reliable than counting prompts, because one long agent run can equal a week of small edits.
Two buying mistakes
- Upgrading for one heavy week rather than a sustained pattern.
- Comparing tiers by model list while ignoring the size of the usage window.
UsageAtlas records reported usage and local session tokens, so the effect of a tier change is measurable rather than guessed. The rolling windows guide explains the timing.
Time the upgrade
Upgrade when blocked days repeat, not when one task is blocked. A single heavy refactor is cheaper to absorb than a month of unused capacity, and a plan change mid-cycle can make the first month look misleadingly good.
- Track blocked days for a full period.
- Note which model was the constraint.
- Compare the price gap with the cost of those days.
- Re-check after the upgrade so the change is measured.
UsageAtlas keeps the before-and-after visible. See rolling windows.
Questions, answered
Is Factory Core better than Standard?
Core includes more capacity and broader access, which matters only if you hit the Standard window often. Otherwise it is unused headroom.
Do both plans use rolling windows?
Usage is measured over windows on both, with different sizes. Check the current plan page for the exact periods.
Can I compare Factory with Cursor?
Yes, but compare cost per completed task. Request counts and window sizes are not directly comparable across tools.