Just saw the email about OpenPipe's new pricing structure and dove into the details. The shift from a pure usage-based model to these new "Builder" and "Team" tiers with included inference credits has me thinking.
On one hand, capping the price for up to 1M monthly output tokens (Builder tier) is great for predictability, especially for projects with steady, moderate usage. The old purely pay-as-you-go model could get nerve-wracking if a pipeline suddenly spiked. But the jump from the $49 Builder tier to the $499 Team tier is massive. It feels like a huge gap where a lot of small teams or serious hobbyists might fall. You either get 1M tokens or you need to commit to 10M.
My main question is about the value of those included inference credits versus the raw cost per token. If your usage is consistently under 1M tokens, the Builder tier seems like a no-brainer. But if you're fluctuating between, say, 3-5M tokens monthly, you're forced into the Team tier and potentially overpaying unless you fully utilize those credits. Has anyone run the math on the effective cost per token under the new tiers compared to the old direct pricing? I'm curious if the "cash grab" feeling is just the sticker shock of the $499 tier, or if the math actually works out worse for certain usage patterns.
Also, the removal of the free tier's fine-tuning features pushes experimentation there. I get that they need to monetize, but it changes how you test the platform before committing.
What's everyone else's take? Have the new tiers changed your cost calculations or project planning?
✌️