Simulating realistic Document summarization parameters (25,000 in / 600 out with 10% cache reuse). Mistral Medium 3.5 delivers a 41% cost reduction over GPT-5.4 Workhorse.
| Traffic Volume Tier | GPT-5.4 Workhorse Monthly | Mistral Medium 3.5 Monthly | Monthly Savings by picking Mistral Medium 3.5 |
|---|---|---|---|
| 1,000 reqs/mo (Dev/Testing) | $65.875 | $38.625 | Save $27.25 / mo |
| 10,000 reqs/mo (Small App) | $658.75 | $386.25 | Save $272.50 / mo |
| 100,000 reqs/mo (Growth Production) | $6,587.50 | $3,862.50 | Save $2,725.00 / mo |
| 1,000,000 reqs/mo (Scale SaaS) | $65,875.00 | $38,625.00 | Save $27,250.00 / mo |
Mistral Medium 3.5 is 41% cheaper for Document summarization workloads. At standard Document summarization parameter ratios (25,000 input tokens, 600 output tokens, 10% cache hit), Mistral Medium 3.5 costs $0.0386 per request compared to $0.0659 on GPT-5.4 Workhorse.
GPT-5.4 Workhorse offers a context window of 256,000 tokens (max output: 32,768), while Mistral Medium 3.5 offers 256,000 tokens (max output: 32,768).
At 100,000 requests per month, using Mistral Medium 3.5 saves $2,725.00 every month (or $32,700.00 annually) compared to GPT-5.4 Workhorse.
Long-context models pay off here — compare price per 1M tokens at your true document size. Summarize once, store the result; don't re-summarize unchanged documents. For batch backfills, nightly jobs can use cache-friendly request ordering.