2024 Update (Forgot the Month)

Ranking

  1. GPT-4o: by OpenAI/Microsoft, Creator of ChatGPT
  2. Claude: by Anthropic, excels in programming, especially within Cursor.
  3. Qwen: by Alibaba
  4. Llama: by Meta
  5. Gemini: by Google
  6. Other

Other large models both domestically and internationally have generally average user experiences, including Musk’s Grok, Douyin’s Duaobao, Kimi, and Wenxin Yiyan, which have ordinary computational and comprehension capabilities.

Summary

  1. The paid version of ChatGPT is the best, but it’s costly, and you can’t just pay to use it. You also need to be cautious about IP stability to avoid using dirty lines or frequently changing IPs, which might lead to restrictions.
  2. The paid version of Claude is also good, and registration requires a foreign phone number. The free version is also acceptable.
  3. Tongyi Qianwen is the best in China.

2025 May Update

Time has come to May 2025.

  • DeepSeek is now well-known with excellent performance. In Chinese environments, if you’re unsure which one to choose, DeepSeek would be a solid option.
  • Google Gemini performs very well and has made significant progress. You can try it out more, and the official website is https://gemini.google.com/
  • ChatGPT has been somewhat average. It used to be the leader, but now it’s just still in the top.

2025 November Update

Overseas

  1. Claude’s code capabilities are probably the leading.
  2. Google Gemini also performs very well.
  3. Grok excels in text creation.
  4. ChatGPT is not impressive, but still stands out.

Domestic

  1. Tencent Yuanbao seems to have made little progress and remains at DeepSeek’s level.
  2. Douyin Duaobao has made rapid progress with frequent updates, but the client is likely based on a frontend ecosystem, a browser shell, which causes noticeable lag in single conversations due to excessive data.
  3. Alibaba Qianwen feels very good when used personally, but seems many people don’t know about it? The Qwen series open-source large models have excellent reputation and applications worldwide, worth expecting.

2025 December Update

Overseas

  1. Claude, Gemini, and ChatGPT are in the first tier. Gemini 3 Pro and ChatGPT 5 alternate in programming development fields.
  2. Grok’s strength seems to be text creation?

Domestic

  1. Tencent Yuanbao, which is integrated with DeepSeek, maintains DeepSeek’s level.
  2. Douyin Duaobao was previously highly praised. It handles daily life issues reasonably, but performs poorly on professional technical questions.
  3. Alibaba Qianwen feels very good when used personally, but seems many people don’t know about it? The Qwen series open-source large models have excellent reputation and applications worldwide, worth expecting.