Musk: Grok Trails Anthropic AI, xAI Needs Time to Catch Up

Elon Musk, the founder of xAI, has publicly acknowledged that his company’s AI model, Grok, is currently outmatched by the latest offerings from rival Anthropic. This admission marks a rare moment of apparent humility from Musk, who is typically known for his unwavering confidence.

The significance of this statement is amplified by the fact that Musk went beyond simply stating Grok needs improvement. He conceded that he had misjudged Anthropic’s position in the artificial intelligence landscape, declaring the company the undisputed leader with no current competitors capable of matching its most advanced models.

xAI launched Grok 4.5 on July 8, 2026, with the intention of making a significant competitive stride. The model was priced aggressively, at $2 per million input tokens and $6 per million output tokens, aiming to present a cost-effective alternative to established AI giants.

However, benchmark results indicate that Grok 4.5 falls short of both Anthropic’s Claude Fable 5 and Opus 4.8 across several key areas. The disparity is particularly evident in coding tasks, where Grok 4.5 achieved a score of only 53% on the DeepSWE 1.1 benchmark.

Musk attributed this performance gap to a straightforward reason: xAI has been operational for approximately three years, whereas Anthropic has had six years to develop its team, refine its research processes, and iterate on its model architectures.

A Complex Relationship: Competition and Collaboration

Despite the competitive tension, Musk’s relationship with Anthropic is not solely adversarial. SpaceX has entered into a substantial compute agreement with Anthropic, securing hundreds of megawatts of power from xAI’s extensive Colossus data center facilities.

Musk has also committed to ensuring equitable access to AI infrastructure, indicating he will not leverage his hardware advantages to disadvantage competitors.

Musk’s Strategy for Advancement

To bridge the performance gap, Musk is pursuing a strategy of deep integration across his various companies. He is actively promoting Grok’s adoption within Tesla and SpaceX, incorporating the models into operational workflows. However, teams at both companies will retain the autonomy to select more effective AI models when necessary.

Musk has expressed optimism that xAI will achieve frontier-level AI performance by 2027. The primary challenge, however, is that Anthropic is not static. The company’s Opus 5.5, which Musk specifically identified as superior to Grok, represents a continually advancing benchmark.