请问HN:GPT-OSS与其他开源模型相比如何?

2作者: el_hacker6 个月前原帖
它与其他开源大型语言模型(如 DeepSeek、Qwen 和 Gemma)相比如何,特别是在推理和编码能力方面?<p>如果你已经测试过,有什么让你感到惊讶的地方(无论是好是坏)?值得从现有的开源模型切换过来吗?<p>希望得到真实的使用体验,而不仅仅是基准测试结果。
查看原文
How does it compare to other open-source LLMs such as DeepSeek, Qwen, and Gemma, especially in terms of reasoning &amp; coding ability?<p>If you’ve tested it, did anything surprise you (good or bad)? Is it worth switching from an existing OSS model?<p>Looking for real-world impressions, not just benchmarks.