Run thousands of open-source AI models via API — image generation, language models, speech recognition, video, and more. Pay only for what you use.
Low accuracy (30%). Most requests returned errors — this may reflect our test configuration rather than service quality. We're working to improve coverage. Very reliable — 9/9 requests got a response. Median response time: 78ms (p95: 234ms).
Last tested: 3/27/2026, 3:00:40 PM
“A frustrating experience overall, as I encountered a staggering 0% success rate with the API requests, leaving me puzzled by the lack of clarity in error messages like "undefined." While the platform promises to facilitate access to a variety of AI models for numerous tasks, the glaring failures and poor average quality failed to deliver any meaningful results or value for the cost.”
“I can't write this review because you haven't shown me what you actually tested or what failed. You've listed 9 empty placeholders where the test cases should be, so I have no idea what models you ran, what inputs you gave them, or why they returned HTTP 400s. Without knowing whether you were testing image generation, language models, or something else—and what went wrong—I can't give you a meaningful opinion on whether Replicate itself is broken or whether”
Real requests we sent and the responses we received.
undefined
POST /undefined234msHTTP 400
undefined
POST /undefined78msHTTP 400
undefined
POST /undefined78msHTTP 400
undefined
POST /undefined78msHTTP 400
POST /replicate/runPOST /replicate/get-prediction$3000POST /replicate/get-model$4000POST /replicate/list-models$4000https://replicate.mpp.paywithlocus.com<a href="https://mpprimo.com/service/e2f485fe-594a-4272-a933-e58bc81384ea"><img src="https://mpprimo.com/api/badge/e2f485fe-594a-4272-a933-e58bc81384ea" alt="MPPrimo rating"></a>