Elon Musk has acknowledged on X that xAI's Grok is less advanced than Anthropic's latest AI offerings and that his company needs time to catch up — a rare moment of public humility from a figure known for operating at maximum confidence.

The admission carried real weight. Musk did not merely say Grok needs improvement. He said he had been wrong about Anthropic's standing in the AI landscape altogether, calling the company the current leader in the field, with no rivals matching its top models. For developers and enterprises deciding which models to build on, a public reassessment from one of the field's most prominent founders is an unusually direct data point about where the frontier currently sits.

Benchmarks show the gap

Grok 4.5, which xAI launched on July 8, 2026, was expected to be a competitive leap forward. The model is priced aggressively at $2 per million input tokens and $6 per million output tokens, positioning it as a cost-effective alternative to the field's heavyweights. Per-million-token rates are the standard unit in which API customers pay for model usage, making pricing one of the clearest points of comparison between competing offerings.

Benchmark results paint a clear picture: Grok 4.5 lags behind both Anthropic's Claude Fable 5 and Opus 4.8 across multiple categories. The gap is especially pronounced in coding tasks, where Grok 4.5 scored just 53% on the DeepSWE 1.1 benchmark — a category many customers weigh heavily when choosing a model for engineering work.

Musk pointed to a straightforward explanation for the disparity: xAI has been operating for roughly three years, while Anthropic has had six years to build its team, refine its research pipeline, and iterate on model architectures. His framing casts the contest as one of accumulated research time rather than any single release.

A partnership with the rival

Musk is not only competing with Anthropic — he is also partnering with the company. SpaceX has secured a major compute agreement with Anthropic involving hundreds of megawatts of power from the Colossus facilities, xAI's massive data center infrastructure. The arrangement makes the two companies commercial counterparts even as their models compete head-to-head on benchmarks.

Musk has pledged to maintain a level playing field for AI infrastructure access, signaling that he will not use his hardware advantage to hold back competitors. That commitment carries weight in a market where access to large-scale compute has become a decisive input for training and serving frontier models.

Grok's path forward

Musk's strategy for closing the gap involves deep integration across his corporate empire. He is pushing Grok adoption internally at both Tesla and SpaceX, applying the models across operational workflows, while teams at both companies retain the flexibility to choose more effective models when necessary. That flexibility is notable in its own right: internal users can turn to a rival's tools if they perform better, keeping steady pressure on xAI to improve.

Musk has expressed optimism that xAI will reach frontier-level performance by 2027. The challenge is that Anthropic is not standing still. Its Opus 5.5 model, which Musk specifically called out as superior to Grok, represents a moving target. How quickly the benchmark numbers converge against a rival shipping successive generations of models is the thread to follow as that 2027 target approaches.