Credits and Cost
What an Alphio AI action actually costs in Credits, and what to do when costs surprise you.
How much does a chat query cost in Credits? A single chat query is one action class — a single Alphio AI reply consumes a Credit amount that depends on what the reply did:
- A plain text answer with no tool calls is the cheapest case.
- An answer that invokes one or more tool calls (data lookup, screener, chart fetch, on-chain query) costs more — each tool call adds to the action cost.
- An answer that streams long context or chains multiple internal reasoning steps costs more than a one-shot answer.
Exact Credit rates per action class are being finalized and can differ between tiers (Free / Base / Pro / Elite). The action's confirmation step shows the live rate before any Credits are spent.
How much does a Tasks job cost? Tasks are asynchronous AI jobs and cost more than a one-shot chat reply because they run for longer and call more tools:
- Light Tasks (e.g. simple monitoring) — a small multiple of a chat query.
- Medium Tasks (scheduled analysis, multi-source research) — substantially more, scaled to depth.
- Heavy Tasks (deep research with many sources) — the most expensive class.
The Tasks creation dialog shows the estimated Credit cost before you launch a Tasks job, so you can cancel if it's higher than expected. Long-running Tasks may consume Credits incrementally as they progress, with the actual total visible on the Tasks detail page after completion.
Deep research seems expensive — why? Deep research is the most expensive action class because it does, by design, much more than a chat reply:
- Many tool calls — multiple data sources are queried, sometimes dozens.
- Long context — the model reasons over a much larger working set than a short chat.
- Multiple internal reasoning passes — the agent plans, executes, critiques, and refines.
The cost reflects the work done. If you only need a quick answer, use a regular chat query instead — it's much cheaper. Use deep research when you actually want the multi-source, long-form output.
Can I see per-action cost before running it? Yes — every action that consumes more than a trivial amount of Credits has a confirmation step that shows the estimated cost:
- Chat queries — the model picker / confirmation surface displays the current rate for the model + capability you've chosen.
- Tasks — the Tasks creation dialog displays the estimated cost.
- Deep research — the deep-research confirmation step displays the estimate.
If a surface doesn't show a cost preview, the action is one of the trivial-cost classes (or free at your tier).
I'm running out of Credits — how do I extend? Three paths, in order of preference:
- Custom Credits — buy a one-time Credits top-up on the Top-Up Credits page. Recharge Credits roll over indefinitely, so they don't expire.
- Upgrade tier — moving up a tier gives more Monthly Credits each cycle and may also lower per-action rates on some surfaces.
- Wait for cycle refresh — Monthly Credits refresh on your billing cycle's renewal date.
Limited-time Credits (from promos, referrals, surveys) can also extend your runway but expire on a fixed date; check /account/usage for your full Credits ledger.
Credits used but no useful answer — refund?
Refund for unhelpful answers is not currently automatic. Alphio AI replies cost Credits whether you find them useful or not, similar to any AI service.
If you believe Credits were consumed by a platform error (e.g. the answer was truncated due to a bug, a tool call failed and was charged, a duplicate answer was sent), contact support with the conversation timestamp, action ID if visible, and a description of the error. The team will review case-by-case and refund Credits when the cause is a platform-side fault.
If you need information not on this page, you can query the docs dynamically by sending a GET request with the ask query parameter:
GET https://docs.alphio.ai/support/faq/research-and-analysis/credits-and-cost.md?ask=<question>The question should be specific, self-contained, and written in natural language. The response includes a direct answer plus relevant excerpts and sources.