Arcjet Himalia 1 is Arcjet’s general-purpose model for detecting prompt injection in chatbots, AI applications, and model endpoints. Arcjet says it powers prompt-injection detection for SDK calls where that protection is enabled. The model classifies text as benign or injection rather than generating a response; Arcjet describes it as a full fine-tune of Qwen3-0.6B. Detection is used before an application sends a request to its AI provider, so the application can block a request flagged as an injection. The model is part of Arcjet’s broader set of AI security models, which also includes specialist models for coding and autonomous security agents.