en.Wedoany.com Reported - On July 8, local time, US artificial intelligence company SpaceXAI released its new AI model, Grok 4.5. Developed in collaboration with AI coding startup Cursor, the model is designed for code writing, agent tasks, and knowledge work scenarios. Its focus is not on ordinary chat-based Q&A, but on enabling the model to handle more complex workflows such as software development, engineering collaboration, legal text processing, and financial services analysis.
Positioned as one of SpaceXAI's most powerful models to date, Grok 4.5 is now available on Cursor. Developers can invoke the model on Cursor's desktop, web, iOS, CLI, and SDK for code generation, codebase understanding, debugging, refactoring, task planning, and multi-step execution. Unlike single-response models, Grok 4.5 emphasizes sustained task processing capabilities, including reading context, breaking down goals, calling tools, modifying code, and providing feedback on execution results.
In this collaboration, Cursor provides a clear entry point for developer scenarios. If an AI coding model only possesses general language capabilities, it struggles to truly complete complex engineering tasks; it also needs to understand code repository structures, dependencies, file calls, test feedback, error logs, and developer instructions. By integrating Grok 4.5 with Cursor, the model can directly enter the development environment rather than remaining confined to an external Q&A window.
In terms of pricing, Cursor disclosed that the input price for the Grok 4.5 base model is $2 per million tokens, and the output price is $6 per million tokens. A faster version is also available, with an input price of $4 per million tokens and an output price of $18 per million tokens. This pricing is clearly aimed at high-frequency engineering invocation scenarios, particularly the demand for large context inputs in code generation, agent execution, and enterprise knowledge work.
SpaceXAI has also listed legal and financial services as target application scenarios. Tasks such as contract review, clause comparison, case retrieval, and compliance document organization in legal services, as well as report generation, data analysis, investment research support, and risk control material processing in financial services, are all long-text, multi-step, and highly context-dependent tasks. For Grok 4.5 to enter these scenarios, it must handle structured documents, reasoning chains, fact-checking, and permission boundaries beyond its coding capabilities.
This release also continues the trend of AI models transitioning from chatbots to work-oriented agents. With Grok 4.5 integrated into Cursor, the model's usage entry points are no longer limited to standalone web pages but are embedded in development tools, command lines, mobile devices, and SDKs. Future deployment will focus on model response speed, code task success rates, complex project context handling, enterprise knowledge base integration, adaptation to legal and financial scenarios, and agent execution stability.
