At 4 requests/user/day (totaling 60,000 queries/mo), running Document summarization on Llama 3.3 70B Instruct costs $291.24 / month ($0.5825 / user / mo).
Document summarization for a 500 active users (early growth) audience means 2,000 requests per day on average, each carrying 25,000 input and 600 output tokens at a 10% cacheable share — so the effective per-request cost on Llama 3.3 70B Instruct is $0.004854, of which $0.00/mo comes back from prompt caching. Per active user that's $0.5825/mo — the number to hold against your pricing tier and gross margin. Charging at least $2.912/user/mo keeps gross margin at or above 80%.
Quick answer
For 500 Active Users, Document summarization on Llama 3.3 70B Instruct is modeled at 60,000 requests per month, costing $291.24 per month or $0.5825 per active user.
Method & trust
This projection combines the selected active-user tier, requests per user per day, input/output tokens per request, and the modeled cache share. It is an operating estimate, not a provider invoice; production traffic and retries should be measured against it.
Total expected API invoice for 500 Active Users generating 60,000 queries.
Direct inference cost per MAU to model your SaaS pricing tiers and gross margins.
Savings generated by exploiting 10% cache hits on prompt context.
| Model | Provider | Monthly Spend (500 Active Users) | Cost / User / Mo | Annual Spend | Context Limit |
|---|---|---|---|---|---|
| Llama 3.3 70B Instruct (Current) | meta | $291.24 | $0.5825 | $3,494.88 | 128,000 |
| GPT-5.6 Terra | openai | $3,162.00 | $6.324 | $37,944.00 | 1,050,000 |
| GPT-5.4 Workhorse | openai | $3,952.50 | $7.905 | $47,430.00 | 256,000 |
| o3-mini | openai | $1,725.90 | $3.452 | $20,710.80 | 200,000 |
| o4-mini | openai | $1,684.65 | $3.369 | $20,215.80 | 256,000 |
| o1-mini | openai | $1,725.90 | $3.452 | $20,710.80 | 128,000 |
At 4 queries per user/day with Llama 3.3 70B Instruct, the estimated cost is $0.5825 per monthly active user (MAU). Total monthly bill for 500 Active Users is $291.24.
Assuming a 10% cache hit rate on repeated system and context tokens, prompt caching saves $0.00 per month ($0.00/year) on Llama 3.3 70B Instruct.
With a direct COGS of $0.5825 per user/month on Llama 3.3 70B Instruct, charging at least $2.912 per user/month ensures an 80%+ software gross margin.