Qwen will be the king?

查看原文

QwenDeepSeek模型发布开源代码

qwen

热度趋势

趋势数据积累中

百分比基于当前可用热度信号，而非评论数或独立用户人数。

推荐理由

Qwen 相关模型动态已经出现，适合跟踪能力变化、生态影响和后续可用性。

AI 摘要

DeepSeek、Qwen 和 GLM 模型正通过扩展推理和后期训练来提升性能，并利用更高的 token 计数。即将发布的 Qwen 4，尽管其是否开源尚不明确，但预计将进一步推动这一趋势。未来整合了“engrams”的模型有望在特定任务上与 2.4T 参数模型匹敌或超越其能力。

Extended reasoning and post-training appear to be the keys used by DeepSeek, Qwen, and GLM to boost performance (leveraging higher token counts). And Qwen 4 hasn't even been released yet. Of course, we don't know if that release will be open-sourced, but I am optimistic about future models, featuring "engrams", that could soon match or surpass 2.4T parameter models on specific tasks.