V-Benchmark aims to be the gold standard for evaluating video AI. We use trained human researchers, automated quality checks coupled with automated AI auditing of scores to ensure we're delivering consistent and accurate scores. We're regularly publishing new evaluations of the frontier models. You can see our methodology here: https://megaton.ai/methedology Our published reports: https://megaton.ai/evaluations/
hi everyone!
After some feedback and iterating we've rebuilt our 2.0 evaluation methodology and launched v in order to create a new gold standard for evaluating creative AI models. We'd really love your feedback on the new look, new site as well as any feedback or thoughts about our methodology and reporting.
Appreciate you all!
Megaton Mask