At 8 requests/user/day (totaling 1,200,000 queries/mo), running Agentic workflow on Llama 4 Scout (109B MoE) costs $8,010.00 / month ($1.602 / user / mo).
Total expected API invoice for 5,000 Active Users generating 1,200,000 queries.
Direct inference cost per MAU to model your SaaS pricing tiers and gross margins.
Savings generated by exploiting 65% cache hits on prompt context.
| Model | Provider | Monthly Spend (5,000 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| Llama 4 Scout (109B MoE) (Current) | meta | $8,010.00 | $1.602 | $96,120.00 | 10,000,000 |
| GPT-5.6 Terra | openai | $61,440.00 | $12.288 | $737,280.00 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $76,800.00 | $15.36 | $921,600.00 | 256,000 |
| o3-mini | openai | $43,560.00 | $8.712 | $522,720.00 | 200,000 |
| o4-mini | openai | $34,980.00 | $6.996 | $419,760.00 | 256,000 |
| o1-mini | openai | $43,560.00 | $8.712 | $522,720.00 | 128,000 |
At 8 queries per user/day with Llama 4 Scout (109B MoE), the estimated cost is $1.602 per monthly active user (MAU). Total monthly bill for 5,000 Active Users is $8,010.00.
Assuming a 65% cache hit rate on repeated system and context tokens, prompt caching saves $0.00 per month ($0.00/year) on Llama 4 Scout (109B MoE).
With a direct COGS of $1.602 per user/month on Llama 4 Scout (109B MoE), charging at least $8.01 per user/month ensures an 80%+ software gross margin.