SpaceXAI has unveiled Grok 4.5, a new AI model tailored for complex coding, legal, and financial tasks. This marks the company’s first product since merging with AI startup Cursor and going public, directly targeting rivals OpenAI and Anthropic with competitive speed and pricing.
- Grok 4.5 beats some rivals on speed, efficiency, and cost
- Targets coding, legal, financial, and cybersecurity tasks
- Priced significantly below Anthropic’s Opus 4.8 and OpenAI’s GPT-5.6
What happened
SpaceXAI launched Grok 4.5, a new AI model developed collaboratively with the recently acquired AI coding startup Cursor. This release represents SpaceXAI’s first major product since going public and prioritizes coding and agentic applications over casual chat. Elon Musk highlighted the model’s efficiency, speed, and cost advantages in announcements on social media.
The model was designed to handle complex, long-duration tasks such as software development, legal analysis, financial operations, and cybersecurity. Grok 4.5 reportedly surpasses Anthropic’s Opus 4.8 on several performance benchmarks while coming in at a lower price per token, making it a competitive offering for enterprise users keen on cost control.
Why it matters
This launch signals SpaceXAI’s strategic pivot toward capturing market share among business and Wall Street clients who demand powerful AI tools for critical workflows. The inclusion of cybersecurity features and expanded vertical focus marks an ambitious push beyond coding alone, addressing sectors traditionally underserved by earlier models.
Cost competitiveness is central to Grok 4.5’s appeal. By pricing input tokens at $2 per million and output tokens at $6, SpaceXAI undercuts Anthropic’s Opus 4.8 and places itself alongside OpenAI’s pricing, but with added speed and efficiency benefits. This pricing aims to attract companies closely managing AI usage costs amid tightening budgets and regulatory scrutiny.
What to watch next
SpaceXAI plans a wider public release of Grok 4.5 soon, though the EU market is not included initially. Given the model uses the same leased compute infrastructure that hosts competitors like Anthropic and Google, SpaceXAI faces choices on whether to expand usage or monetize those resources differently as demand grows.
Meanwhile, OpenAI is rolling out its GPT-5.6 with new voice features, amid regulatory pressures to pace launches carefully. The AI sector remains highly competitive and closely watched by governments, especially around cybersecurity safeguards. How Grok 4.5 performs in real-world adoption and its ability to close the gap to the largest models could set the tone for SpaceXAI’s future growth.