Classification: CALCULATED illustrative scenario using editable inputs—not observed workload spend or a plan recommendation.
View the full infographic → · View this article
A monthly API estimate is reproducible only when request volume, current and target token counts, cache share, output usage, model, rates and currency conversion are visible. The number is not a bill unless those inputs match real traffic and every charge category is included.
One calculator scenario
The Context Cost Calculator’s illustrative GPT-6 Luna scenario uses 10,000 requests per month, 12,000 current input tokens per request, 8,000 target input tokens, 50% and 40% cached-input shares, and zero output tokens entered for both sides. It uses displayed USD rates of $0.10 uncached input, $0.01 cached input and $0.50 output per million tokens. Zero output is an example input, not an observed workload count.
With those assumptions and the ECB reference conversion dated 1 October 2026, the calculator displays about £4.9873 current and £3.8689 target, a calculated difference of £1.1184. This result depends on the example’s request count, cache shares, price preset and FX reference. It is not observed spend or a guaranteed saving.
Check the boundaries
The scenario excludes possible cache-write or storage treatment, tool/grounding fees, taxes, discounts and quality/retry effects. Provider price pages can change; recheck the official source and verification date. Compare accepted work rather than token volume alone.
Use the calculator to change inputs and view price and ECB sources. Subscription mode is different: the £89 Pro 100 charge is user-recorded and the benchmark is in progress, so no per-token subscription cost is implied.