The cost calculus for small builders is shifting faster than I expected. A few months ago, using a cheaper Chinese model felt like a tradeoff: you saved money but got noticeably worse output. That gap is closing, and in some cases it has closed entirely. I've been running the same prompts through DeepSeek and a couple others against what I was using before, and the difference for practical tasks like summarizing customer feedback, drafting copy, and generating boilerplate is small enough that I'm having a hard time justifying the price difference.
The harder part to reason about is trust and data handling. For a hobbyist project it barely matters. For anything touching user data it matters a lot, and the answers there are murky.
What I keep coming back to is that the cost compression is happening at the model layer, and that changes the math for anyone building on top of these APIs.
Curious whether people here have actually switched any of their regular workflows over, or are still treating the cheaper options as secondtier.
[link] [comments]