At 10 requests/user/day (totaling 15,000 queries/mo), running RAG / search-augmented answers on Microsoft Phi-4 (14B) costs $14.25 / month ($0.285 / user / mo).
Total expected API invoice for 50 Active Users generating 15,000 queries.
Direct inference cost per MAU to model your SaaS pricing tiers and gross margins.
Savings generated by exploiting 50% cache hits on prompt context.
| Model | Provider | Monthly Spend (50 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| Microsoft Phi-4 (14B) (Current) | microsoft | $14.25 | $0.285 | $171.00 | 16,384 |
| GPT-5.6 Luna | openai | $22.20 | $0.444 | $266.40 | 1,050,000 |
| GPT-5.4 mini | openai | $83.25 | $1.665 | $999.00 | 256,000 |
| GPT-5.4 nano | openai | $22.575 | $0.4515 | $270.90 | 128,000 |
| GPT-4o mini | openai | $18.00 | $0.36 | $216.00 | 128,000 |
| Claude Haiku 4.5 | anthropic | $103.50 | $2.07 | $1,242.00 | 1,000,000 |
At 10 queries per user/day with Microsoft Phi-4 (14B), the estimated cost is $0.285 per monthly active user (MAU). Total monthly bill for 50 Active Users is $14.25.
Assuming a 50% cache hit rate on repeated system and context tokens, prompt caching saves $0.00 per month ($0.00/year) on Microsoft Phi-4 (14B).
With a direct COGS of $0.285 per user/month on Microsoft Phi-4 (14B), charging at least $1.425 per user/month ensures an 80%+ software gross margin.