The invention provides an optimal
service provider large model selection method and device for an API
system, and relates to the technical field of
artificial intelligence, and the method comprises the steps: carrying out the fusing filtering, concurrent hard
current limiting filtering and weighted random routing filtering of an initial
service provider large model set according to model identification data and state data, and obtaining a preferred
service provider large model; calling the first-choice service provider large model to a
server of the API
system, and obtaining a calling state of the first-choice service provider large model; when the calling state is calling failure, rejecting the preferred service provider large model from the initial service provider large
model set, constructing a degraded service provider large
model set based on the remaining service provider large models, and performing fusing filtering, concurrent hard
current limiting filtering and weighted random routing filtering on the degraded service provider large
model set to obtain a degraded service provider large model set; re-screening a first-choice service provider large model; when the calling state is calling success, taking the preferred service provider large model as an optimal service provider large model; and generating an answer text of the
question text through the optimal service provider large model.