AI engineering teams spend weeks finding, running, scaling, and collaborating on scattered AI evaluations/benchmarks. The solution is Benchmarkthing.com — Evals as an API (Think Replicate/HuggingFace for evals and agents starting with evals). Benchmarkthing was founded by Xiangyi and Moritz, who spent months building evals for Tesla, WeChat Video, Princeton researchers, and several YC companies. Within 2 weeks of inception, we generated more than $4k in revenue with 7 paying users. Both Xiangyi and Moritz are repeat founders. Xiangyi used to work at Tesla, Red Hat, and Dolby, and Moritz raised $2.8M from General Catalyst for RemNote. We’re raising a pre-seed round. Recently filled half of the round after several checks from AI infra experts and tier-1 scout fund.
About us: @xdotli @moritzw42
Contact: xiangyi@benchmarkthing.com