At 10 requests/user/day (totaling 750,000 queries/mo), running RAG / search-augmented answers on Pixtral Large costs $8,850.00 / month ($3.54 / user / mo).
Total expected API invoice for 2,500 Active Users generating 750,000 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 (2,500 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| Pixtral Large (Current) | mistral | $8,850.00 | $3.54 | $106,200.00 | 128,000 |
| GPT-5.6 Sol | openai | $27,750.00 | $11.10 | $333,000.00 | 1,050,000 |
| GPT-5.6 Cyber | openai | $69,375.00 | $27.75 | $832,500.00 | 1,050,000 |
| GPT-5.5 Standard | openai | $27,750.00 | $11.10 | $333,000.00 | 512,000 |
| GPT-5.5 Pro | openai | $166,500.00 | $66.60 | $1,998,000.00 | 512,000 |
| o3-pro (Frontier Reasoning) | openai | $96,000.00 | $38.40 | $1,152,000.00 | 1,000,000 |
At 10 queries per user/day with Pixtral Large, the estimated cost is $3.54 per monthly active user (MAU). Total monthly bill for 2,500 Active Users is $8,850.00.
Assuming a 50% cache hit rate on repeated system and context tokens, prompt caching saves $5,400.00 per month ($64,800.00/year) on Pixtral Large.
With a direct COGS of $3.54 per user/month on Pixtral Large, charging at least $17.70 per user/month ensures an 80%+ software gross margin.