{"version":"1.0","generated_at":"2026-11-22T10:33:52.789773","id":4093,"slug":"new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop","title":"新开源模型 Inkling 2 Mini 通过循环思考用更少文字解决问题，仍可在笔记本电脑上运行","summary":"Thinking Machines 发布开源循环式 Transformer 模型 Inkling 2 Mini。它总参数 26.4B、每 token 激活 5.4B，通过 24 层中 12 层循环三次减少输出 token，支持文本图像和 100 万 token 上下文，采用 Apache 2.0 许可。原文称其在 100B 以下开源模型的智能体编程测试中领先，并可在 32GB Mac 与 Nvidia DGX Spark 本地运行。","abstract":"新开源模型 Inkling 2 Mini 通过循环思考用更少文字解决问题，仍可在笔记本电脑上运行 Thinking Machines 发布开源循环式 Transformer 模型 Inkling 2 Mini。它总参数 26.4B、每 token 激活 5.4B，通过 24 层中 12 层循环三次减少输出 token，支持文本图像和 100 万 token 上下文，采用 Apache 2.0 许可。原文称其在 100B 以下开源模型的智能体编程测试中领先，并可在 32GB Mac 与 Nvidia DGX Spark 本地运行。 Thinking Machines 发布开源循环式 Transformer 模型 Inkling 2 Mini。它总参数 26.4B、每 token 激活 5.4B，通过 24 层中 12 层循环三次减少输出 token，支持文本图像和 100 万 token 上下文，采用 Apache 2.0 许可。原文称其在 100B 以下开源模型的智能体编程测试中领先，并可在 32GB Mac 与 Nvidia DGX Spark 本地运行。 原贴提到：Thinking Machines has released Inkling 2 Mini, a so-called looped transf 来源：the-decoder.com Thinking Machines 发布了 Inkling 2 Mini，一种所谓的循环式 Transformer。 循环式 Transformer 会为每个 token 多次运行某些层。 Inkling 2 Mini 将其 24…","access_level":"member","access_label":"免费阅读","access_mode":"full","is_preview":false,"canonical_url":"https://opc.beizhux.com/content/4093/new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop","html_url":"https://opc.beizhux.com/content/4093/new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop","json_url":"https://opc.beizhux.com/content/4093/new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop.json","published_at":"2026-11-21T18:33:02","updated_at":"2026-11-22T10:29:03","category":{"slug":"ai-business","name":"AI小生意项目库"},"source":{"site":"the-decoder.com","author":"Matthias Bastian","url":"https://the-decoder.com/new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop/"},"tags":["MoE","开源模型","循环Transformer","循环式Transformer","推理成本","推理效率","智能体编程","混合专家","端侧AI"],"topics":[{"slug":"ai-side-business","name":"AI副业","url":"https://opc.beizhux.com/topics/ai-side-business","reason":"这篇内容命中「项目、小生意、变现」等主题信号。"},{"slug":"agent-workflow","name":"Agent工作流","url":"https://opc.beizhux.com/topics/agent-workflow","reason":"这篇内容命中「智能体」等主题信号。"},{"slug":"ai-tools","name":"AI工具","url":"https://opc.beizhux.com/topics/ai-tools","reason":"这篇内容命中「模型」等主题信号。"}],"keywords":["AI小生意项目库","AI副业","Agent工作流","AI工具","工具","自动化","模型","Cursor","Claude","Agent","智能体","工作流"],"questions":[{"question":"新开源模型 Inkling 2 Mini 通过循环思考用更少文字解决问题，仍可在笔记本电脑上运行主要讲什么？","answer":"Thinking Machines 发布开源循环式 Transformer 模型 Inkling 2 Mini。它总参数 26.4B、每 token 激活 5.4B，通过 24 层中 12 层循环三次减少输出 token，支持文本图像和 100 万 token 上下文，采用 Apache 2.0 许可。原文称其在 100B 以下开源模型的智能体编程测试中领…"},{"question":"这篇文章最值得关注的要点是什么？","answer":"Thinking Machines 发布开源循环式 Transformer 模型 Inkling 2 Mini。它总参数 26.4B、每 token 激活 5.4B，通过 24 层中 12 层循环三次减少输出 token，支持文本图像和…；原贴提到：Thinking Machines has released Inkling 2 Mini, a so-called looped transf；来源：the-decoder.com"},{"question":"这篇文章和哪些AI专题相关？","answer":"它适合放在AI副业、Agent工作流、AI工具专题里阅读。 关联原因：这篇内容命中「项目、小生意、变现」等主题信号。；这篇内容命中「智能体」等主题信号。；这篇内容命中「模型」等主题信号。"},{"question":"阅读这篇文章建议先理解哪些关键词？","answer":"建议先理解AI工具、工具、自动化、模型、Cursor这些关键词，再结合正文判断工具、机会或风险是否值得进入自己的工作流。"}],"terms":[{"slug":"ai-tools-term","name":"AI工具","definition":"在AI觉醒星球里，「AI工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-tools-term","path":"/glossary/ai-tools-term","json_url":"https://opc.beizhux.com/glossary/ai-tools-term.json"},{"slug":"tools","name":"工具","definition":"在AI觉醒星球里，「工具」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/tools","path":"/glossary/tools","json_url":"https://opc.beizhux.com/glossary/tools.json"},{"slug":"automation","name":"自动化","definition":"在AI觉醒星球里，「自动化」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/automation","path":"/glossary/automation","json_url":"https://opc.beizhux.com/glossary/automation.json"},{"slug":"ai-model","name":"模型","definition":"在AI觉醒星球里，「模型」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/ai-model","path":"/glossary/ai-model","json_url":"https://opc.beizhux.com/glossary/ai-model.json"},{"slug":"cursor","name":"Cursor","definition":"在AI觉醒星球里，「Cursor」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/cursor","path":"/glossary/cursor","json_url":"https://opc.beizhux.com/glossary/cursor.json"},{"slug":"claude","name":"Claude","definition":"在AI觉醒星球里，「Claude」属于「AI工具」方向。围绕AI工具、模型能力、自动化流程和真实使用场景做整理，优先关注能提升效率、降低成本和创造新交付的工具。 不只看工具热度，更看它能不能进入真实流程。","topic_slug":"ai-tools","topic_name":"AI工具","topic_title":"AI工具：模型、插件、自动化与实战场景","topic_url":"https://opc.beizhux.com/topics/ai-tools","topic_path":"/topics/ai-tools","url":"https://opc.beizhux.com/glossary/claude","path":"/glossary/claude","json_url":"https://opc.beizhux.com/glossary/claude.json"},{"slug":"agent","name":"Agent","definition":"在AI觉醒星球里，「Agent」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/agent","path":"/glossary/agent","json_url":"https://opc.beizhux.com/glossary/agent.json"},{"slug":"ai-agent","name":"智能体","definition":"在AI觉醒星球里，「智能体」属于「Agent工作流」方向。聚合Agent、AI编程、Claude Code、Cursor、工作流自动化等内容，帮助你把智能体能力落到真实任务链路。 把智能体当成任务系统，而不是聊天窗口。","topic_slug":"agent-workflow","topic_name":"Agent工作流","topic_title":"Agent工作流：智能体、AI编程与自动化协作","topic_url":"https://opc.beizhux.com/topics/agent-workflow","topic_path":"/topics/agent-workflow","url":"https://opc.beizhux.com/glossary/ai-agent","path":"/glossary/ai-agent","json_url":"https://opc.beizhux.com/glossary/ai-agent.json"}],"preview_hidden_sections":[],"related_items":[{"id":4096,"title":"两个英伟达的故事","url":"https://opc.beizhux.com/content/4096/a-tale-of-two-nvidias","summary":"TechCrunch Equity 播客讨论英伟达本周的“两个叙事”：财报营收飙升、现任管理层称无需黄仁勋回归；同时马斯克等人呼吁黄仁勋回归，且公司停止在中国销售 H200 芯片并收紧远程访问限制。播客还提到批评者估计约 20% 业务来自中国，但非财报直接披露。","access_level":"member"},{"id":4089,"title":"英伟达研究者认为，超大 AI 上下文窗口是条死路","url":"https://opc.beizhux.com/content/4089/nvidia-researchers-argue-huge-ai-context-windows-are-a-dead-end","summary":"英伟达研究团队指出，让模型一次吃下整本小说往往既昂贵又没必要；在多组测试中，定期摘要加定向检索的效果差不多。文章同时回顾了上下文窗口从 GPT-4 的 8,192 tokens 到今天约一百万 tokens 的膨胀过程。","access_level":"member"},{"id":4097,"title":"Transformers v5.24.0 发布：新增多款模型与架构支持","url":"https://opc.beizhux.com/content/4097/release-v5-24-0","summary":"Hugging Face Transformers v5.24.0 发布说明，新增 Inkling 2 Mini、Apertus 1.5、Dory、HY4-VL、Omnilingual ASR、Lara Mini、SarvamMLA、SarvamMoE、EXAONE 等模型支持，并包含 WavTokenizer、模型…","access_level":"member"}],"usage_policy":{"summarizable":true,"preferred_url":"https://opc.beizhux.com/content/4093/new-open-model-inkling-2-mini-thinks-in-loops-to-solve-problems-with-fewer-words-and-still-runs-on-a-laptop","private_fields_excluded":true}}