Automatic Model Selection

The technology works by automatically choosing among available AI models based on the task at hand, rather than routing every request to a single model. According to Selectstar, this selective routing results in a 28% reduction in cost compared with fixed model usage.

The company presented the technology publicly, marking its formal introduction to the market. Details on which specific models the system selects among, and how the selection criteria are weighted, were not disclosed in the announcement.

Cost Savings as the Core Claim

The 28% figure is the central quantitative claim attached to the announcement. Savings of this kind would arise from directing simpler requests to cheaper, smaller models while reserving larger, more expensive models for tasks that require them — a common rationale in model-routing approaches, though the company's own methodology for measuring the 28% saving was not detailed in the material made public.

Selectstar's move comes as Korean enterprises face rising inference costs tied to LLM deployment, making cost-optimization layers around AI models an area of active commercial interest.

What Remains Unclear

The announcement leaves open several questions: how the 28% saving was benchmarked, against what baseline workload, and whether the figure reflects internal testing or customer deployments. No independent validation of the cost-reduction claim has been reported, and the company's disclosure does not specify the deployment environments in which the technology has been tested.

For now, the technology stands as a company claim: an auto-selection system for AI models that Selectstar says cuts costs by 28%. Given that the technology exists as announced and the savings figure is the company's own uncontradicted measurement, the claim is Mostly True.