qwen3 VS deepseek

[ironkingdom2017-beep]

1. Why are only the two large models Qwen3 and DeepSeek adopted, while other models cannot be applied to the agent-enabled internet?
2. Each model has its own token limit. What should we do if the word count of the response exceeds the token limit?