Elon Musk's Grok 4.5 model leads the world ranking for complex tasks

A significant turning point has occurred in the AI technology race. The new Grok 4.5 model from xAI, founded by Elon Musk, has taken first place in the Long-Horizon Terminal-Bench (LHTB) ranking for performing complex logical tasks. This result means that the ability of AI agents to not only process text but also solve multi-step problems has reached a new level. This is reported by Ixbt.com reports .

Long-Horizon Terminal-Bench is a specialized platform that tests AI models in a terminal (command line) environment. It requires the system to perform long-term planning, analysis, and sequential decision-making. Grok 4.5 recorded an average score of 0.505 in these tests, successfully completing 13 out of 46 highly complex tasks.

Outpacing competitors and technical growth

The significance of this achievement is that Grok 4.5 has surpassed the most powerful models on the market. Specifically, it is claiming the status of the smartest AI agent, outperforming giants like Claude Sonnet 5, Opus 4.8, Fable 5, and even OpenAI's GPT-5.5. This indicates that the xAI team has made a massive leap in algorithm optimization in a short time.

For comparison, the previous version of the Grok model, 4.20, was unable to complete a single task within this benchmark. The journey from zero results to global leadership in just a few months shows that Elon Musk's massive investments in computing power (NVIDIA chips and supercomputers) are paying off.

Future possibilities and significance

The success of the Grok 4.5 model is not just about numbers, but a step toward creating assistants capable of replacing humans in practical programming and system management. The ability to work in a terminal environment allows AI agents to perform complex server operations, independently fix bugs in software code, and manage large databases.

This news is naturally of great interest to users and developers in Uzbekistan. Although Grok models are currently primarily provided through the X (formerly Twitter) platform, its API capabilities could become a strong foundation for local startups and technological solutions in the future. Elon Musk stated that this model is not only fast but also closer to human reasoning than ever before.

According to ixbt.com, this ranking, compiled at the end of July 2026, has shifted the balance of power in the AI world. Now, other technology corporations, including Google and Microsoft, will have to take more serious measures to increase the "competence" level of their models.