What Does An AI Trading Strategy Actually Cost To Run?

Anywhere from 2 cents a day to several dollars an hour. Here's exactly how evaluation costs add up on Everstrike, and how to keep them low.

What Does An AI Trading Strategy Actually Cost To Run?

One of the most common questions we get about vibetrading is: "How much will this cost me?"

The honest answer is: anywhere from 2 cents a day to several dollars an hour. It depends almost entirely on three choices you make when you set up your strategy.

This post breaks down exactly how costs add up on Everstrike, what the free credits cover, and how to keep your bill small.

What you pay for

An AI strategy doesn't cost anything to create. It costs money to run.

Every time a strategy runs, the AI looks at the market, reads its data and decides whether to act. That's called an evaluation (or a check). Each evaluation costs a small amount, depending on the model.

With the default model, GPT-6 Luna, an evaluation costs $0.0005.

That's it. There are no trading fees on Everstrike, no subscription and no charge for placing orders. You pay for evaluations, and nothing else.

πŸ’‘
Only strategy evaluations cost credits. Backtests in the Strategy Builder run in your browser and are free.

The formula

Each check costs the model's price once for the strategy itself, plus once for every market it covers:

Cost per Check = Model Price x (1 + Number of Markets)

So a one-market strategy on GPT-6 Luna costs $0.0005 x 2 = $0.001 per check. A three-market strategy costs $0.0005 x 4 = $0.002 per check.

Then multiply by how often it runs:

Cost per Day = Cost per Check x Checks per Day

The three choices that decide your bill

1. Evaluation frequency. This is by far the biggest one. A strategy that checks every 30 seconds runs 2,880 times a day. One that checks every hour runs 24 times.

2. Number of markets. Every market adds the model's price again to each check.

3. Model. GPT-6 Luna is the default and costs $0.0005 per check. Some models cost more. Very fast frequencies (15 seconds or less) use GPT-4.1-nano, which is cheaper per check but runs far more often.

Here's what a one-market strategy on GPT-6 Luna costs per day:

Evaluation frequencyChecks per dayCost per day
Every hour24$0.024
Every 15 minutes96$0.096
Every 5 minutes288$0.29
Every minute1,440$1.44
Every 30 seconds2,880$2.88
Bar chart of daily cost for a one-market strategy, from $0.024 checking hourly to $2.88 every 30 seconds, with the $0.25 free daily refill marked.
Daily cost of a one-market strategy on the default model. Everything left of the green line fits inside the free daily refill.
πŸ’‘
The evaluation frequency is the biggest lever on cost. Match it to the timeframe your strategy actually trades.

Free credits

Every Everstrike account starts with free AI credits for completing a few onboarding steps.

On top of that, until you've deposited more than $4.50, your free credit balance refills to $0.25 every day at 00:00 UTC.

What does $0.25 a day buy? From the table above, a one-market strategy that checks every hour costs about $0.024 a day. That's roughly ten such strategies running every day for free - or one strategy checking every 15 minutes, with room to spare.

Once you've deposited more than $4.50 on Mainnet, your USD balance pays for evaluations instead (1 credit is $1), and the free daily refill ends. Deposits made in DEX mode and CEX mode both count.

What happens when you run out

Your strategies keep running as long as your balance covers one evaluation of every running strategy.

If it drops below that, your strategies show No credits and stop running. Nothing is lost - they resume automatically as soon as your balance is sufficient again, either from the next daily refill or from a deposit.

You can also pause a strategy at any time. Paused strategies don't evaluate, don't trade and don't cost anything.

A credit gauge with a marker for one check of every running strategy: above it strategies run, below it they show No credits.
Strategies run while your balance covers one check of each of them, and resume automatically when it does again.

How to keep costs low

  • Lower the evaluation frequency. A strategy that trades the 4h timeframe gains almost nothing from checking every 30 seconds. The candles it reads only change every four hours.
  • Start with one market. Add more once the strategy has proven itself.
  • Limit the input data. By default, every evaluation sends a full set of data to the model. Removing data your prompt doesn't use lowers the cost per evaluation.
  • Turn on Cost Savings Mode. It reduces token consumption by up to 30%, at the cost of shorter reasoning on each event.
  • Keep the lookback short. The lookback period (5 bars by default) sets how much history each evaluation reads. More history means more tokens.
  • Pause what you're not using. A paused strategy costs nothing.

Where to see what you're spending

Before you submit a strategy, the Preview tab shows its expected cost.

Once it's running, the AI credits page shows your available credits, what each strategy used over the last 7, 30 or 90 days, the daily cost of everything that's running, and how long your balance will last at that pace.

Three tiles for the three cost levers - frequency, markets and model - with the cost-per-check formula below.
Three choices decide the bill. Frequency matters most.

Wrapping up

  • You pay per evaluation. With GPT-6 Luna, that's $0.0005 per check, once for the strategy plus once per market.
  • A one-market strategy costs about 2 cents a day at an hourly frequency, and $2.88 a day checking every 30 seconds.
  • Until you've deposited more than $4.50, free credits refill to $0.25 every day at 00:00 UTC.
  • The cheapest strategy is the one that checks only as often as it needs to.
VIBETRADING

See what your strategies cost on the AI credits page.

Open AI credits

Subscribe to Everstrike

Don’t miss out on the latest issues. Sign up now to get access to the library of members-only issues.
jamie@example.com
Subscribe