Cached vs Uncached OpenAI Cost

See how prompt caching saves money. Adjust the sliders to estimate your savings.

Model

System message tokens: 10,000

User input tokens (per request): 500

Output tokens (per request): 300

Total requests: 1,000

Cache hit rate (%): 70

Without Caching

$57.00

for 1,000 requests

With Caching

$25.50

Save $31.50 (55.3%)

System message: $18.50 (with 70% cache hit)

User input: $2.50 (never cached)

Output: $4.50