SpaceXAI Launches Grok 4.5, Its Most Advanced AI Model for Coding and Business Tasks

By: Harsh Vardhan

|

Updated on: 09-Jul-2026 04:00 PM

9,078 views

Follow Us:

9,078 views

SpaceXAI has launched Grok 4.5, its most advanced AI model to date, aiming to compete with leading AI labs Anthropic and OpenAI. The new model, built in partnership with the coding platform Cursor, is designed to handle complex coding, agentic tasks, and business applications such as finance and legal services.

Key Highlights

• SpaceXAI launches Grok 4.5 in partnership with Cursor for advanced coding and business tasks
• Grok 4.5 outperforms Anthropic’s Opus 4.8 in DeepSWE 1.0 benchmark and is just behind GPT-5.5
• Model trained on tens of thousands of Nvidia GB300 GPUs with enhanced cybersecurity features
• Grok 4.5 priced lower than Claude Opus 4.8 and GPT 5.6 Terra for both input and output tokens

Grok 4.5 Features and Capabilities

Grok 4.5 is the first joint model from SpaceXAI and Cursor. The launch follows SpaceX's recent acquisition agreement with Cursor. According to SpaceXAI, Grok 4.5 is capable of managing difficult, long-running tasks, especially in software engineering. It also extends its capabilities to legal and financial work, matching the offerings of Anthropic’s Claude models.

Elon Musk, founder of SpaceXAI, claims Grok 4.5 can match Anthropic’s Claude Opus 4.7 model but operates at a much faster speed. Musk emphasized that Grok 4.5 combines strong capabilities, faster speed, and lower cost, making it competitive in the AI market.

The model is now the default in Grok Build and is available through Cursor on all plans and the SpaceXAI console. It can create complex Excel models using web research and multi-sheet formulas, leave notes for later reference, build diagrams in PowerPoint, and write clear prose in Word. SpaceXAI demonstrated Grok 4.5 by generating a Three.js solar system simulation with adjustable time and realistic motion.

Performance and Security Enhancements

SpaceXAI shared benchmark results for Grok 4.5. In the DeepSWE 1.0 AI coding benchmark, Grok 4.5 outperformed Anthropic’s Opus 4.8 and ranked just behind OpenAI’s GPT-5.5. The model was trained using tens of thousands of Nvidia GB300 GPUs, highlighting a significant investment in infrastructure.

Grok 4.5 also includes enhanced cybersecurity features. Cursor stated it has implemented measures to detect and block malicious actors while preserving legitimate security research, such as finding and patching vulnerabilities. These steps aim to restrict workflows that could cause harm, addressing concerns that previously led the US government to limit access to advanced AI models like Claude Fable 5 and GPT-5.6. Both models have since received approval for general release.

Market Position and Pricing

The release of Grok 4.5 is part of SpaceXAI’s broader strategy to catch up in the AI race and attract more business customers. Earlier this year, Musk acknowledged that xAI, now merged with SpaceXAI, had fallen behind in coding capabilities, prompting staffing changes. In May, SpaceXAI released its first coding agent to compete with Anthropic’s products.

Grok 4.5 is priced at $2 (about Rs 190) per million input tokens and $6 (about Rs 570) per million output tokens. This is lower than Claude Opus 4.8, which costs $5 (about Rs 480) and $25 (about Rs 2,390) respectively. OpenAI’s GPT 5.6 Terra is priced at $2.5 (about Rs 240) for input and $15 (about Rs 1,430) for output tokens. Grok 4.5 is not yet available in the European Union.