Budget tiers
Parallect's six research budget tiers -- caps, default provider counts, and how actual billing works.
Budget tiers
A budget tier is the maximum amount a single research job is allowed to spend. It also influences how many providers participate by default and how much depth you can comfortably ask for.
You are billed the actual cost of the run (within your account's pricing rules), not the tier maximum by default -- the tier is a ceiling, not a fixed fee.
The six tiers
| Tier | Approx. cap | Default providers | What you get |
|---|---|---|---|
| XXS | $1 | 1 | Minimal spend; one focused pass -- best for tiny checks. |
| XS | $2 | 1 | Light research; one strong provider pass. |
| S | $5 | 1 | Standard "answer my question well" without going huge. |
| M | $15 | 2 | Comfortable depth for most professional questions. |
| L | $30 | 3 | Broad coverage across models and angles. |
| XL | $60 | All | Maximum breadth -- every available provider in play. |
You can override the default provider selection to include more or fewer providers within your tier's budget.
Budget Tier Selector
Click a tier to see what you get
How to choose the right tier
Ask three questions:
- How wrong can we afford to be? -- If the answer guides spend, compliance, or safety, favor M+ and often methodical mode.
- How many independent lenses do we need? -- One strong lens (XXS-S) vs. corroboration and disagreement (M-XL).
- Is this exploratory? -- Use XXS-XS for brainstorming and step up once the question stabilizes.
Practical defaults:
- Quick factual or internal note -- XXS or XS.
- Typical business or product question -- S or M.
- Board deck, legal-ish scan, or multi-stakeholder decision -- L or XL with multiple providers.
Cost examples in practice
These are illustrative -- real jobs vary with prompt length, mode, and provider behavior.
- Fast mode, single provider, short question -- Often lands well under a few dollars; an XXS or XS cap is plenty if you picked an economical provider.
- Methodical mode, two providers + synthesis -- Commonly sits in the single digits to low teens of dollars for many queries at M tier -- still below the cap unless the question explodes in scope.
- Methodical, "all hands" with premium models -- Can approach higher dollar amounts; L or XL caps protect you while allowing the run to complete.
If you're unsure, start one tier lower than you think, run fast once, then escalate tier and mode when you see what's missing.
How budget caps and billing relate
- Cap -- The tier sets the maximum customer cost Parallect will allow for that job.
- Actual charge -- When the job finishes, you pay what it cost (including synthesis), up to the cap.
- Why caps exist -- They prevent runaway spend if a query grows unexpectedly large or a provider works harder than anticipated.
The tier protects you from unexpectedly high costs. You're only charged what the job actually costs, up to the cap.
Related reading
- Creating research threads -- where you pick tier and see cost preview.
- Research modes -- how fast (parallel) and methodical (sequential) affect your research.