[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"$fPy4pOICTF9snyeHp4-LO-3ZgWZTbT6_-ai_jHX8Y98c":3},{"code":4,"msg":5,"data":6},200,"操作成功",{"id":7,"title":8,"content":9,"digest":10,"source":10,"coverPath":11,"thumbsCoverPath":12,"isTop":13,"isShow":14,"baseClick":13,"clickCount":15,"createTime":16,"typeId":17,"isNewest":18,"newsInfoTypeRespVo":19,"voiceUrl":22,"voiceSize":23,"taskId":24,"releaseTime":25,"titleEn":26,"contentEn":27,"voiceUrlEn":28,"taskIdEn":29,"voiceSizeEn":30},1333,"DeepSeek新大招曝光：下一步智能体","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">DeepSeek下一步，被曝剑指智能体。\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">知情人士透露，DeepSeek正在开发具有更强大AI Agent能力的新模型，预计在今年年底就会推出。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F072947ca96144f79b93e42900ed93742\u002Fdadcd4d23c7b48a08f1f33b7c42fd299.webp\" width=\"682\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cstrong style=\"color: rgb(187, 187, 187);\" class=\"ql-lineHeight-1-75\">△\u003C\u002Fstrong>\u003Cspan style=\"color: rgb(187, 187, 187);\" class=\"ql-lineHeight-1-75\">图片由AI生成\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-center\">\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">官方尚无回应，但这一爆料并非无迹可寻。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">在DeepSeek 8月的最近一次更新中，DeepSeek-V3.1上线。官方发布的升级公告中就提到：\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek-V3.1具备\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">更强的Agent能力\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">，通过Post-Training优化，新模型在工具使用与智能体任务中的表现有较大提升。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">迈向Agent时代的第一步\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F2cd93889b74f49028ffb224aa339d7bf\u002F9668c1d5487e48219b2f16c9bc1a90da.webp\" width=\"775\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\">“R2”将造更强智能体？\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">彭博社援引知情人士消息，爆料DeepSeek的下一个模型，仅需少量提示，就能帮用户执行复杂操作，并且还能根据历史操作自我进化和学习。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">这一模型将在今年最后一个季度面世。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">此前，业内普遍推测，按照DeepSeek V3和DeepSeek R1的发布节奏，在被期待已久的“R2”之前，DeepSeek应该会先一步推出新一代V系列的基础模型。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">而从去年12月到今年8月的近9个月时间里，DeepSeek V系列的版本号只从V3来到了V3.1。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">看上去仅仅是小版本的改进，但从DeepSeek官方后来公布的测试结果来看，似乎又暗藏玄机。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek-V3.1的主要升级包括3点：\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">混合推理架构，就是把思考模式和非思考模式都集成到了一个模型里。\u003C\u002Fstrong>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">更高的思考效率，比DeepSeek-R1想得更快。\u003C\u002Fstrong>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">更强的Agent能力，新模型在工具使用与智能体任务中的表现有较大提升。\u003C\u002Fspan>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">具体来看提升效果。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">编程智能体\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">方面，在代码修复测评SWE和命令行终端环境下的复杂任务（Terminal-Bench）测试中，DeepSeek-V3.1相较V3和R1有明显提高。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fee7676318e1747519bb6dd4c65642ca7\u002F3474887f4d36462fadb905ffc91dd693.webp\" width=\"666\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">而在\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">搜索智能体\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">测评中，相较于DeepSeek-R1-0528，DeepSeek-V3.1也展现出了全面的性能提升。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F28a9781afec04c20be748509cf13878e\u002Fad00a67b4047497188afe950debcd98c.webp\" width=\"664\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\">下一个DeepSeek时刻\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek R1将推理模型推向发展新高潮，成了搅动全球大模型行业格局的一尾鲶鱼。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Manus让智能体的概念火爆出圈，被称为“第二个DeepSeek时刻”——\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">来到2025年的下半年，几乎没有大模型产品不谈智能体。吴恩达也在最新Andrew’s Letters中提到：并行智能体正在成为提升AI能力的新方向。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fde7c20c5c31343b780d810dac64fee3a\u002F931b8eb1104f45ecb82b756e82c13ac8.webp\" width=\"673\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">早早预定2025年年度热词的DeepSeek和智能体一叠加，立刻又触动了人们兴奋的神经。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">已经有人开始期待：如果是DeepSeek来做这件事，那智能体的价格门槛，真的要被打下来了。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">但也有人保持怀疑：这个时候入局智能体，还算好时机吗？\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">不管怎么说，可以确定的是，在2025年余下的最后1\u002F3时间里，大模型领域的热闹依然将每天持续。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">最后的最后，小小One More Thing一下，信源来自官方，妥妥靠谱：\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Qwen 3 max也要来了吗？\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">欢迎一起，见证下一个历史时刻。\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">【新闻来源】量子位 | 公众号 QbitAI \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.sohu.com\u002Fa\u002F932056748_610300?scm=10001.1104_13-100000-0_922.0-0.0.a2_5X162X1653&spm=smpc.channel_159.block3_218_AB1PKt_1_fd.3.1757070156519ACWzA26_1104\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002Fx1yxZ\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（本网转发此文章，旨在为读者提供更多的信息资讯，所涉内容不构成投资、消费建议。文章事实如有疑问，请与有关方核实，文章观点非本网观点，仅供读者参考。）\u003C\u002Fspan>\u003C\u002Fp>","","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F712c0ddfa0d142b19bc4d33a6a1befff\u002FAI领域.jpg","https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fthumbs\u002F712c0ddfa0d142b19bc4d33a6a1befff\u002FAI领域.jpg",0,1,52,"2025-09-05 19:07",2,false,{"id":17,"name":20,"enName":21},"芯位视野","Xinwei Vision","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3A7bd11d35-2055-4433-9af4-0bdce4fccf4d%3A0.wav?Expires=1757082822&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=ppXfrkFFqaRfmAUJQxjxoG%2FwhpQ%3D",5706990,"7bd11d35-2055-4433-9af4-0bdce4fccf4d","2025-09-05 19:02","DeepSeek's New Big Move Revealed: Next-Generation AI Agent","\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">DeepSeek's Next Step, Exposed to Target AI Agents.\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Insiders revealed that DeepSeek is developing a new model with even stronger AI agent capabilities, expected to be launched by the end of this year.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F072947ca96144f79b93e42900ed93742\u002Fdadcd4d23c7b48a08f1f33b7c42fd299.webp\" width=\"682\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp class=\"ql-align-center\">\u003Cstrong style=\"color: rgb(187, 187, 187);\" class=\"ql-lineHeight-1-75\">△\u003C\u002Fstrong>\u003Cspan style=\"color: rgb(187, 187, 187);\" class=\"ql-lineHeight-1-75\">Image generated by AI\u003C\u002Fspan>\u003C\u002Fp>\u003Cp class=\"ql-align-center\">\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Official has not responded yet, but this leak is not without evidence.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In DeepSeek's recent update in August, DeepSeek-V3.1 was launched. The official upgrade announcement mentioned:\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek-V3.1 has\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Stronger Agent Capabilities\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">, with significant improvements in tool usage and agent tasks through Post-Training optimization.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong class=\"ql-lineHeight-1-75\" style=\"font-size: 18px; color: rgb(255, 153, 0);\">The First Step Toward the Agent Era\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F2cd93889b74f49028ffb224aa339d7bf\u002F9668c1d5487e48219b2f16c9bc1a90da.webp\" width=\"775\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\">Will \"R2\" Create Stronger AI Agents?\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Bloomberg cited insider information, revealing that DeepSeek's next model can assist users in performing complex operations with just a few hints, and it can also self-evolve and learn based on historical operations.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">This model will be released in the last quarter of this year.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Previously, the industry generally speculated that, following the release rhythm of DeepSeek V3 and DeepSeek R1, DeepSeek should first launch a new base model of the V series before the highly anticipated \"R2\".\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">However, over nearly nine months from December last year to August this year, the version number of the DeepSeek V series only advanced from V3 to V3.1.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">It seems like just a minor version update, but according to the test results later released by DeepSeek's official, there might be more to it.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">The main upgrades of DeepSeek-V3.1 include three points:\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Hybrid reasoning architecture, which integrates both thinking mode and non-thinking mode into one model.\u003C\u002Fstrong>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Higher reasoning efficiency, thinking faster than DeepSeek-R1.\u003C\u002Fstrong>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cul>\u003Cli>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Stronger Agent capabilities, with significant improvements in tool usage and agent tasks in the new model.\u003C\u002Fspan>\u003C\u002Fli>\u003C\u002Ful>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Looking at the specific improvement effects.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Programming agents\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">, in code repair evaluation SWE and complex tasks in command line terminal environments (Terminal-Bench), DeepSeek-V3.1 shows significant improvements compared to V3 and R1.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fee7676318e1747519bb6dd4c65642ca7\u002F3474887f4d36462fadb905ffc91dd693.webp\" width=\"666\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">In the\u003C\u002Fspan>\u003Cstrong style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">search agent\u003C\u002Fstrong>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\"> evaluation, DeepSeek-V3.1 also showed comprehensive performance improvements compared to DeepSeek-R1-0528.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002F28a9781afec04c20be748509cf13878e\u002Fad00a67b4047497188afe950debcd98c.webp\" width=\"664\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cstrong style=\"font-size: 18px; color: rgb(255, 153, 0);\">The Next DeepSeek Moment\u003C\u002Fstrong>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek R1 pushed the inference model to a new peak, becoming a catfish that stirred up the global large model industry landscape.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Manus made the concept of AI agents go viral, being called the \"second DeepSeek moment\" —\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">By the second half of 2025, almost no large model product fails to mention AI agents. Andrew Ng also mentioned in his latest Andrew’s Letters: Parallel AI agents are becoming a new direction for enhancing AI capabilities.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cimg alt=\"undefined\" src=\"https:\u002F\u002Fimage.51xinwei.com\u002F2025\u002F09\u002Fde7c20c5c31343b780d810dac64fee3a\u002F931b8eb1104f45ecb82b756e82c13ac8.webp\" width=\"673\" height=\"undefined\" style=\"display: block; margin: auto;\">\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">DeepSeek and AI agents, which were already pre-ordered as the hot words of 2025, instantly triggered people's excitement.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Some have already begun to look forward to: If DeepSeek does this, then the price barrier of AI agents may really be broken down.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">But others remain skeptical: Is it a good time to enter the AI agent field now?\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Regardless, it is certain that in the remaining one-third of 2025, the excitement in the large model field will continue every day.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Finally, a small One More Thing, the source comes from the official, very reliable:\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Is Qwen 3 max coming too?\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cspan style=\"font-size: 18px;\" class=\"ql-lineHeight-1-75\">Welcome to witness the next historical moment together.\u003C\u002Fspan>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp>\u003Cbr>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">[News Source] Quantum Bit | Official Account QbitAI \u003C\u002Fspan>\u003Ca href=\"https:\u002F\u002Fwww.sohu.com\u002Fa\u002F932056748_610300?scm=10001.1104_13-100000-0_922.0-0.0.a2_5X162X1653&spm=smpc.channel_159.block3_218_AB1PKt_1_fd.3.1757070156519ACWzA26_1104\" rel=\"noopener noreferrer\" target=\"_blank\" style=\"color: rgb(187, 187, 187);\">http:\u002F\u002Fu5a.cn\u002Fx1yxZ\u003C\u002Fa>\u003C\u002Fp>\u003Cp class=\"ql-align-justify\">\u003Cspan style=\"color: rgb(187, 187, 187);\">（This article is reposted by this website to provide readers with more information. The content does not constitute investment or consumption advice. If there are any questions about the facts of the article, please verify with the relevant parties. The views expressed in the article are not the views of this website and are for reference only.）\u003C\u002Fspan>\u003C\u002Fp>","https:\u002F\u002Fxinwei-dev-test.oss-cn-shenzhen.aliyuncs.com\u002Fintelligent\u002Faudio%3Ad8a7f7ef-f104-4a1f-9774-22c956e2d3e4%3A0.wav?Expires=1774838476&OSSAccessKeyId=LTAI5tNvY2RkKjZw4LLWsrPK&Signature=oBzAL8wMOmjvdV1lCT5ChcZbsnk%3D","d8a7f7ef-f104-4a1f-9774-22c956e2d3e4",6462380]