Liquid AI launches RIPL inference engine promising sub-50 ms responses
Liquid AI introduced RIPL, a low-latency inference engine that serves its LFM2.5 models in a live request path. The company claims end-to-end responses under 50 ms for models ranging from 230 million to 2.6 billion parameters. The comparisons come from Liquid's own tests, and the 50 ms target does not hold for longer requests.
