Free tier economics

In brief

If one prompt costs ~$0.005 and one filled impression pays ~$0.0095 at $9.50 CPM, inference-time ads can cover a free tier—the rest is fill rate and volume.

Start with unit costs

Every prompt has marginal cost (tokens, GPU time, provider fees). Use your real average cost per prompt; the example below uses $0.005.

Monthly compute cost = prompts_per_month × cost_per_prompt.

Translate CPM into per-prompt revenue

Programmatic ads are priced in CPM (cost per thousand impressions). Revenue per filled impression = CPM / 1000.

  • Revenue per prompt = (CPM / 1000) × fill_rate × slots_per_prompt

Worked example

Assume 500k prompts/month, 1 slot/prompt, 100% fill, $9.50 CPM, and $0.005 cost/prompt:

  • Compute cost: 500,000 × $0.005 = $2,500
  • Ad revenue: 500,000 / 1000 × $9.50 = $4,750
  • Net: +$2,250

What changes the result

  • Fill rate: no-bid, blocks, and timeouts reduce impressions.
  • eCPM: varies by geo, format, and demand.
  • Slots per prompt: multiple surfaces change inventory.
  • Cost per prompt: depends on model choice and prompt length.
  • Consent rate: semantic matching can require opt-in.

Next step

If you want a broader comparison, see AI monetization. For the operating model, start with How wavebird works.

Related

Last updated: 2026-04-03

Next step

Explore the integration path.

Start with the SDK for the production path, or talk to the team if you want rollout, placement, or partnership guidance.