In the midst of their war against Trump, Claude is already the most valued AI in four of the nine top categories.

According to the aggregated data from LLM Arena, the leading comparison platform for AIs based on evaluations conducted by actual users, Claude, the family of models developed by Anthropic, has established itself as the highest rated in four of the nine main categories that currently define the performance of a language model. This leadership comes at a time of tremendous popularity for Claude models, in which the company led by Dario Amodei is engaged in an open conflict with the Trump Administration, after refusing to comply with its demands regarding the use of its AI for citizen surveillance and autonomous weapons.

At present, Claude leads the rankings for Text, Programming, and Document Analysis. In fact, in the latter category, the capacity of Anthropic’s models to handle massive context windows without losing coherence has displaced solutions previously regarded as unbeatable such as Google’s Gemini or OpenAI’s GPT models. The model Claude Opus 4-6, especially in its version optimized for deep thinking, is the undisputed leader.

Moreover, Claude is, perhaps surprisingly, also leading in the Search category. This is one of the most hotly contested verticals, with Grok (from xAI) climbing the ranks to surpass GPT 5.2, which now falls to third place.

However, Claude’s dominance is not absolute. While Anthropic continues to hold the crown in language and pure logic, Google has managed to safeguard its leadership in the area of Artificial Vision as well as Image and Video Generation from text, with its Gemini and Veo models. Meanwhile, Grok leads in video generation from images, and ChatGPT holds the crown in the image editing category.

If we look at the overall combined ranking, in which the scores from the individual categories are taken into account, we find that two Claude models occupy places among the top three positions:

1. Claude Opus 4-6
2. Claude-opus-4-6-thinking
3. Gemini-3.1-pro
4. Grok-4.20-beta1
5. Gemini-3-pro
6. Gpt-5.4-high
7. Gpt-5.2-chat
8. Gemini-3-flash
9. Grok-4.1-thinking
10. Claude-opus-4-5

As you can see, although OpenAI aims to continue climbing in the rankings with its new GPT-5.4 models, it faces the challenge of not being the leader in the categories that provide the greatest value to the business sector, such as code and reasoning. Only time will tell whether, in the days ahead, its new models surpass those from Anthropic.

Image: Gemini