SpaceXAI Releases Grok 4.5 After Cursor Acquisition: Musk Returns to AI Competition Frontline
On July 9, 2026, SpaceXAI (formerly xAI) officially released its first AI model Grok 4.5 after acquiring Cursor. Elon Musk called this model 'Opus-class' on social media, meaning its performance can rival Anthropic's top-tier models. This release marks Musk's significant strategic shift in the AI field - from previous aggressive competition to resource integration through acquisitions, returning to the first tier of AI competition. Notably, Musk also publicly stated 'I was clearly wrong about Anthropic', acknowledging that Anthropic is the true leader in the current AI field.
The backdrop of Grok 4.5's release is quite complex. SpaceXAI (formerly xAI) has undergone significant changes over the past year. At the end of 2025, the company announced the acquisition of AI code editor Cursor, a move that shocked the entire tech industry. Cursor is known for its excellent AI-assisted coding capabilities and has a large loyal user base. After the acquisition, SpaceXAI gained Cursor's deep technical expertise in code understanding and generation, laying an important foundation for Grok 4.5's development.
From a technical specification perspective, Grok 4.5 adopts a brand-new architecture design. According to information disclosed by SpaceXAI, the model excels in instruction following, agentic tool use, and accuracy. The pricing strategy is also quite aggressive: only $1.25 per million input tokens and $2.50 per million output tokens, significantly lower than OpenAI's GPT-5.6 Sol ($5/$30) and Anthropic's Claude Fable 5 ($10/$50). Additionally, Grok 4.5 supports configurable reasoning capabilities, allowing users to adjust the model's thinking depth based on task complexity, flexibly balancing speed and cost.
A notable feature of Grok 4.5 is its deep integration with Cursor. After acquiring Cursor, SpaceXAI obtained all technical assets of this popular AI code editor. According to insiders, Grok 4.5 inherits Cursor's core advantages in code understanding and generation, particularly excelling in handling large codebases, understanding project context, and generating high-quality code. This makes Grok 4.5 especially suitable for research, enterprise workflows, and general-purpose applications.
Musk's attitude shift toward Anthropic is the most noteworthy aspect of this release. For a long time, Musk's relationship with Anthropic was quite tense. He had publicly criticized Anthropic's 'safety-first' strategy, believing it hindered AI development speed. However, as Anthropic achieved tremendous success in the enterprise market with Claude Code, Musk's attitude underwent a noticeable shift. He stated on social media: 'I was clearly wrong about Anthropic, they are the clear AI leader.' This statement was interpreted by outsiders as a rare acknowledgment of competitors by Musk, also reflecting the profound changes in the AI industry landscape.
Notably, there exists a complex business relationship between SpaceXAI and Anthropic. According to reports, Anthropic pays SpaceXAI $1.25 billion per month in compute costs, making the two companies both competitors and partners. This subtle relationship is not uncommon in the AI industry, but it is staggering in scale. Musk's attitude shift may also be related to this business relationship - acknowledging Anthropic's leading position helps maintain the cooperative relationship between both parties, ensuring SpaceXAI can continue to benefit from this huge transaction.
From a market perspective, Grok 4.5's release further intensifies competition in the AI model market. Currently, the AI market has formed a tripartite structure with OpenAI, Anthropic, and SpaceXAI. OpenAI holds advantages in the consumer market with the GPT-5.6 series, Anthropic has established a strong position in the enterprise market through Claude Code, and SpaceXAI has found its place in the developer tools market through the Cursor acquisition. This competitive landscape is good news for users - more choices mean lower prices and better products.
Grok 4.5's pricing strategy is particularly noteworthy. The $1.25/$2.50 pricing makes it one of the cheapest frontier models currently on the market. This 'low price, high quality' strategy is clearly aimed at rapidly capturing market share. For budget-conscious developers, Grok 4.5 provides an extremely attractive option - achieving near-top-tier performance without bearing high usage costs. This pricing strategy may also force competitors to further reduce prices, thereby benefiting the entire user base.
For developers, Grok 4.5's configurable reasoning capability is an important differentiating feature. Users can adjust the model's thinking depth based on task requirements: for simple tasks, fast mode can be used to save costs; for complex reasoning tasks, deep thinking mode can be enabled for more accurate results. This flexibility enables Grok 4.5 to adapt to various application scenarios, from simple code completion to complex system design. Developers can use Evergreen Tools'API成本计算器 to estimate usage costs under different reasoning modes, making optimal choices. Meanwhile, JSON格式化工具 and 代码压缩工具 can help developers process AI-generated code more efficiently.
FAQ
Q1: What are the main differences between Grok 4.5 and previous Grok models?
A: Grok 4.5 is the first model released after SpaceXAI acquired Cursor, called 'Opus-class' by Musk, meaning its performance reaches top-tier levels. Main differences include: 1) Significant improvements in instruction following, agentic tool use, and accuracy; 2) Configurable reasoning capability, allowing users to adjust thinking depth based on task complexity; 3) Deep integration with Cursor, with significant advantages in code understanding and generation; 4) More aggressive pricing at $1.25/$2.50 per million tokens, making it one of the cheapest frontier models available.
Q2: Why did SpaceXAI acquire Cursor?
A: SpaceXAI's acquisition of Cursor was primarily to gain its technical expertise and user base in the AI code editing field. Cursor is a popular AI code editor with a large loyal user base and deep code understanding technology. Through the acquisition, SpaceXAI can quickly enhance Grok model capabilities in code-related tasks while establishing a strong foothold in the developer tools market. This strategic move enables SpaceXAI to form differentiated advantages in competition with OpenAI and Anthropic.
Q3: Why did Musk say 'I was clearly wrong about Anthropic'?
A: Musk's statement reflects a shift in his understanding of the AI industry landscape. Previously, Musk criticized Anthropic's 'safety-first' strategy as hindering AI development speed. However, as Anthropic achieved tremendous success in the enterprise market with Claude Code, becoming the clear leader in AI, Musk acknowledged his misjudgment. Additionally, there exists a business relationship between SpaceXAI and Anthropic (Anthropic pays SpaceXAI $1.25 billion monthly in compute costs), and acknowledging Anthropic's leading position helps maintain the cooperative relationship between both parties.
Q4: What does Grok 4.5's configurable reasoning capability mean?
A: Configurable reasoning capability means users can adjust the model's thinking depth based on task requirements. For simple tasks (like code completion, simple Q&A), fast mode can be used where the model outputs results quickly, saving cost and time; for complex tasks (like system design, complex reasoning), deep thinking mode can be enabled where the model spends more time 'thinking', providing more accurate and comprehensive answers. This flexibility enables users to make optimal tradeoffs between speed, cost, and accuracy based on specific scenarios.
Q5: What use cases is Grok 4.5 suitable for?
A: Grok 4.5 is particularly suitable for: 1) Code development - excels in code understanding, generation, and refactoring thanks to Cursor integration; 2) Research work - configurable reasoning capability enables handling complex analysis and reasoning tasks; 3) Enterprise workflows - accurate instruction following makes it suitable for automating enterprise processes; 4) General AI applications - low price and high quality make it ideal for budget-conscious users. Developers can use Evergreen Tools'在线代码编辑器 to test Grok 4.5's code capabilities.
Summary
Grok 4.5's release marks a significant turning point for SpaceXAI in the AI competition. Through acquiring Cursor to gain technical expertise, using aggressive pricing strategy to capture market share, and acknowledging Anthropic's leading position to maintain business relationships, Musk has demonstrated a more pragmatic and flexible competitive strategy. For users, Grok 4.5 provides an extremely cost-effective choice, especially in code development and complex reasoning tasks. As AI market competition intensifies, we can expect to see more innovative products and more competitive pricing emerge. Evergreen Tools will continue tracking the latest AI industry developments, providing you with the most comprehensive technical information and practical tools.