Grok 4.7 Model Updates and New Safety Identifier Field

This update to the SpaceXAI release notes details a new feature for associating end-user identifiers with API requests and clarifies the specifications for the Grok 4.7 model. Developers can now include a 'safety_identifier' in various API calls, such as Chat Completions and Batch API requests. This identifier helps SpaceXAI attribute any policy violations to a specific end-user of an application rather than the application's API key. The previous 'user' field remains functional. For the Grok 4.7 model, which is designed for coding, agentic, and knowledge tasks, its 'Fast' variant is further defined. This Grok 4.7 Fast model is noted to cost twice the standard token rates, or 1.5 times for long-context requests. It is exclusively available within Cursor and Grok Build platforms, not via the public SpaceXAI API, and billing is managed through those platforms. The standard Grok 4.7 model continues to offer a 500k context window, accept text and image inputs for text-only output, and has no text output limit. Its pricing structure is $2 per 1M input tokens, $0.50 per 1M cached input tokens, and $6 per 1M output tokens for prompts under 200k tokens, with increased rates for prompts exceeding that threshold.