IdeaSift

Developers struggle to keep LLM cost estimates accurate

Developers comparing or budgeting for LLM APIs need current prices across model versions, but some maintain provider price lists by hand. Basic per-token comparisons and calculators can miss important cost factors, including provider-specific caching and image inputs, making estimates less useful for real workloads.

For LLM application developers. Mentioned from Jun 2024 to Feb 2026 on Hacker News.

5 different people described this problem in 3 separate discussions.

Week of 2026-07-13: 0Week of 2026-07-20: 0Week of 2026-07-27: 0Week of 2026-08-03: 0Week of 2026-08-10: 0Week of 2026-08-17: 0Week of 2026-08-24: 0Week of 2026-08-31: 0Week of 2026-09-07: 0Week of 2026-09-14: 0Week of 2026-09-21: 0Week of 2026-09-28: 0
0 mentions in the last 12 weeks
Indie fit
4.0/10
Pain
5.2/10
Frequency
6.5/10
Willingness to pay
0.0/10
Momentum
5.0/10
Who pays
Businesses
Competition
High
Build difficulty
Medium

What people said

Quoted word for word. Follow a link to read the whole discussion.

  1. Does anyone have an API that maintains a list of all model versions for a provider? I hand-update OpenAI into a JSON file that I use for cost reporting in my apps (and in an npm package called llm-primitives).
    binarymax on Hacker NewsJul 2025Asked for a toolHas a workaround
  2. I find the caching pricing and strategy (since different providers use a different implementation of caching as well as different pricing) to be a major factor rather than just the "raw" per token cost, and that is missing here, as well as on the Simon Willison site [1]
Build brief

See what to build and who will buy it

  • 2 product ideas with the smallest useful version and pricing
  • 4 places to find your first customers
  • 3 more quotes from people who have this problem
  • Current workarounds, existing solutions and risks