At 200 requests/user/day (totaling 60,000 queries/mo), running Data extraction & tagging on QwQ 32B (Reasoner) costs $57.36 / month ($5.736 / user / mo).
Total expected API invoice for 10 Active Users generating 60,000 queries.
Direct inference cost per MAU to model your SaaS pricing tiers and gross margins.
Savings generated by exploiting 20% cache hits on prompt context.
| Model | Provider | Monthly Spend (10 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| QwQ 32B (Reasoner) (Current) | qwen | $57.36 | $5.736 | $688.32 | 128,000 |
| GPT-5.6 Terra | openai | $376.80 | $37.68 | $4,521.60 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $471.00 | $47.10 | $5,652.00 | 256,000 |
| o3-mini | openai | $184.80 | $18.48 | $2,217.60 | 200,000 |
| o4-mini | openai | $178.20 | $17.82 | $2,138.40 | 256,000 |
| o1-mini | openai | $184.80 | $18.48 | $2,217.60 | 128,000 |
At 200 queries per user/day with QwQ 32B (Reasoner), the estimated cost is $5.736 per monthly active user (MAU). Total monthly bill for 10 Active Users is $57.36.
Assuming a 20% cache hit rate on repeated system and context tokens, prompt caching saves $8.64 per month ($103.68/year) on QwQ 32B (Reasoner).
With a direct COGS of $5.736 per user/month on QwQ 32B (Reasoner), charging at least $28.68 per user/month ensures an 80%+ software gross margin.