The following is the answer Gemini gave me after I asked it earlier:
- Claude: The top choice for literary writing and long-context handling
Core strengths:
Claude, especially versions such as Opus and Sonnet 3.5, has consistently ranked highly in several blind-test benchmarks, including LMSYS Chatbot Arena, particularly in areas such as “Creative Writing” and “Instruction Following.” This makes it especially suitable for literary writing, long-form novel planning, and writing tasks that require stable handling of extensive context.
- ChatGPT: Strong logical reasoning and all-purpose work assistance
Core strengths:
ChatGPT, especially GPT-4o, performs reliably in complex logical reasoning, multi-angle analysis, and comprehensive comparisons. For example, it is highly useful when comparing the strengths and weaknesses of DeepSeek and Qwen in web novel plot design, or when assisting with tasks such as information organization, strategic planning, and document polishing.
Can anyone see what the problem is?