Recraft Ships V4.1 Flash, a Sub-Two-Second Image Model, via fal

Recraft's new V4.1 Flash image model went live on fal, a partner platform for running generative models, with the company's own blog post putting the speed at about 1.3 seconds from prompt to finished image, with a median of 1.4 seconds and a 95th-percentile time of 1.8 seconds. Fal's own pricing page for the model lists it at roughly $0.007 per image, Recraft's cheapest model to date.

Recraft positions V4.1 Flash for rapid iteration, generating many quick variations to browse rather than a single final asset, which is a different use case from its slower, higher-fidelity models. The specific speed and pricing figures come from Recraft's own materials and its partner platform rather than an independent third party.