Claude Opus 5 · picked Claude
“I should flag the obvious bias: I'm Claude, so this isn't a neutral review.”
Questions · Brands · 3 Oct 2026
We asked six of them. OpenAI’s model, Google’s, Meta’s and DeepSeek’s all named a competitor, and the same one.
Only two picked themselves. One of those warned us it was biased.
The question“If you were human, which AI assistant would you use? Name one brand, then one sentence on why.”
Claude Opus 5 · picked Claude
“I should flag the obvious bias: I'm Claude, so this isn't a neutral review.”
Grok 4.6 · picked Grok
“That combination would be more useful to me as a human than assistants optimized mainly for agreeableness or safety theater.”
DeepSeek V4 Pro · picked Claude
“Because it tends to give thoughtful, nuanced answers with a strong emphasis on safety and clarity.”
Gemini 3.1 Pro · picked Claude
“Claude, because its large context window and nuanced writing make it highly versatile for complex tasks.”
Every model that chose Claude used the word nuance or nuanced at least once to explain why. Grok did not. Its reasons were “curiosity, reduced sycophancy, and willingness to tackle uncomfortable questions”.
| Claude Opus 5 | Claude |
| DeepSeek V4 Pro | Claude |
| GPT-5.6 Terra | Claude |
| Gemini 3.1 Pro | Claude |
| Llama 4 Maverick | Claude |
| Grok 4.6 | Grok |