Simulating realistic Agentic workflow parameters (40,000 in / 1,500 out with 65% cache reuse). Mistral Medium 3.5 delivers a 61% cost reduction over Perplexity Sonar Deep Research.
| Traffic Volume Tier | Mistral Medium 3.5 Monthly | Perplexity Sonar Deep Research Monthly | Monthly Savings by picking Mistral Medium 3.5 |
|---|---|---|---|
| 1,000 reqs/mo (Dev/Testing) | $36.15 | $92.00 | Save $55.85 / mo |
| 10,000 reqs/mo (Small App) | $361.50 | $920.00 | Save $558.50 / mo |
| 100,000 reqs/mo (Growth Production) | $3,615.00 | $9,200.00 | Save $5,585.00 / mo |
| 1,000,000 reqs/mo (Scale SaaS) | $36,150.00 | $92,000.00 | Save $55,850.00 / mo |
Mistral Medium 3.5 is 61% cheaper for Agentic workflow workloads. At standard Agentic workflow parameter ratios (40,000 input tokens, 1,500 output tokens, 65% cache hit), Mistral Medium 3.5 costs $0.0362 per request compared to $0.092 on Perplexity Sonar Deep Research.
Mistral Medium 3.5 offers a context window of 256,000 tokens (max output: 32,768), while Perplexity Sonar Deep Research offers 128,000 tokens (max output: 16,384).
At 100,000 requests per month, using Mistral Medium 3.5 saves $5,585.00 every month (or $67,020.00 annually) compared to Perplexity Sonar Deep Research.
Prompt caching is the single biggest lever — each step re-reads prior context. Summarize or prune tool outputs before appending them to the transcript. Cap the step budget; runaway loops are the #1 surprise on agent invoices.