At 10 requests/user/day (totaling 30,000,000 queries/mo), running RAG / search-augmented answers on Cerebras — Qwen 3 32B costs $108,000.00 / month ($1.08 / user / mo).
Total expected API invoice for 100,000 Active Users generating 30,000,000 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 (100,000 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| Cerebras — Qwen 3 32B (Current) | cerebras | $108,000.00 | $1.08 | $1,296,000.00 | 128,000 |
| GPT-5.6 Luna | openai | $44,400.00 | $0.444 | $532,800.00 | 1,050,000 |
| GPT-5.4 mini | openai | $166,500.00 | $1.665 | $1,998,000.00 | 256,000 |
| GPT-5.4 nano | openai | $45,150.00 | $0.4515 | $541,800.00 | 128,000 |
| GPT-4o mini | openai | $36,000.00 | $0.36 | $432,000.00 | 128,000 |
| Claude Haiku 4.5 | anthropic | $207,000.00 | $2.07 | $2,484,000.00 | 1,000,000 |
At 10 queries per user/day with Cerebras — Qwen 3 32B, the estimated cost is $1.08 per monthly active user (MAU). Total monthly bill for 100,000 Active Users is $108,000.00.
Assuming a 50% cache hit rate on repeated system and context tokens, prompt caching saves $0.00 per month ($0.00/year) on Cerebras — Qwen 3 32B.
With a direct COGS of $1.08 per user/month on Cerebras — Qwen 3 32B, charging at least $5.40 per user/month ensures an 80%+ software gross margin.