Loading threat actor…
Xingjiabi is a financially motivated large language model access-resale operation that advertises discounted access to frontier AI models and has been linked to the use of stolen cloud credentials. Its observed activity involves validating compromised Amazon Bedrock credentials and discovering the model access and service quotas those credentials provide. In October 2026, infrastructure associated with Xingjiabi checked permissions on two stolen enterprise credentials, enumerated inference profiles, regional model availability, and Bedrock quotas for one credential, and manually tested the other. These checks occurred after high-volume abuse had already begun from unrelated infrastructure. Xingjiabi's involvement in the original credential theft or the subsequent bulk model usage has not been established. The operation also maintained a distillation-behavior detection interface that classified users into risk categories. No country of origin or state sponsorship has been established.
Who, where, and from where this actor operates. Pulled from open-source reporting and Mallory's analyst review.
Sectors the actor has been observed targeting.
3 indicators attributed to this actor. Values are masked here and available in full in Mallory.
2 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.