At 40 requests/user/day (totaling 120,000,000 queries/mo), running AI coding assistant on DeepInfra — Llama 3.3 70B costs $216,000.00 / month ($2.16 / user / mo).
Total expected API invoice for 100,000 Active Users generating 120,000,000 queries.
Direct inference cost per MAU to model your SaaS pricing tiers and gross margins.
Savings generated by exploiting 60% cache hits on prompt context.
| Model | Provider | Monthly Spend (100,000 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| DeepInfra — Llama 3.3 70B (Current) | deepinfra | $216,000.00 | $2.16 | $2,592,000.00 | 128,000 |
| GPT-5.6 Terra | openai | $4,204,800.00 | $42.048 | $50,457,600.00 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $5,256,000.00 | $52.56 | $63,072,000.00 | 256,000 |
| o3-mini | openai | $2,164,800.00 | $21.648 | $25,977,600.00 | 200,000 |
| o4-mini | openai | $1,927,200.00 | $19.272 | $23,126,400.00 | 256,000 |
| o1-mini | openai | $2,164,800.00 | $21.648 | $25,977,600.00 | 128,000 |
At 40 queries per user/day with DeepInfra — Llama 3.3 70B, the estimated cost is $2.16 per monthly active user (MAU). Total monthly bill for 100,000 Active Users is $216,000.00.
Assuming a 60% cache hit rate on repeated system and context tokens, prompt caching saves $0.00 per month ($0.00/year) on DeepInfra — Llama 3.3 70B.
With a direct COGS of $2.16 per user/month on DeepInfra — Llama 3.3 70B, charging at least $10.80 per user/month ensures an 80%+ software gross margin.