OLLAMA CLOUD, UNLEASHED

Plan-worth calculator

Enter your traffic. See what each request costs at a model’s listed prices, and whether the included credits last the month.

Verified 12 Oct 2026

Sources: ollama.com/pricing

Your traffic

Standard prices from the pricing page, verified 12 Oct 2026. Default: glm-5.3-flash, the model that dominated our 8 Oct 2026 calibration mix.

Default 345: 1.47B input tokens in 30 days, divided by 142,000 per request, is about 10,350 requests, or 345 a day.

Default 142,000: the agent-heavy account average.

Default 93%: the cached share of that account's 30-day input.

Default 1,200: the same account's average.

glm-5.3-flash has no off-peak row, so this share is not applied.

What it costs

Per request
$0.006053
Per day
$2.09
Per 30 days
$62.65
Runs out

Pro: $60 included

The included credits run out after about 29 days, on day 29 of a 30-day month.

Fits

Max: $300 included

Thirty days use $62.65 of the $300 included credits.

Your $20 Pro subscription

Pro costs $20 a month and includes $60 of usage credits. At these rates those credits cover about 9,913 requests a month, or about 330 a day. Purchased credits are extra.

Prices verified 12 Oct 2026. Source: ollama.com/pricing. Credits: Pro $60 and Max $300 a month, from the same page.

The formula

fresh input per request  = input tokens x (1 - cached share)
cached input per request = input tokens x cached share
cost per request = (fresh x input rate + cached x cached rate + output x output rate) / 1,000,000
rate = (1 - off-peak share) x standard rate + off-peak share x off-peak rate   (off-peak rows only)
per day = cost per request x requests per day
per 30 days = per day x 30
days until Pro credits run out = $60 / per day     (Max: $300 / per day)
requests on $60 = $60 / cost per request

Everything runs in your browser. Nothing you type is sent anywhere. Estimates use the list prices only: real bills can differ by rounding, retries and purchased credits.

Check the real numbers

The calculator uses list prices and your own averages. The checker reads your real usage from your key, and nothing is stored.

Check your own usage