[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-hunyuan-gander-omni-interaction-agent":3,"topics-all":38,"news-related-089f56f3-32ff-4036-89b5-728d5f5a9359":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"089f56f3-32ff-4036-89b5-728d5f5a9359","边聊边干活:腾讯混元开源全模态交互 Agent Gander,小脑管对话、大脑管执行","腾讯混元团队开源 Gander:基于 MiniCPM-o 4.5 的「小脑」负责实时全双工对话,可插拔的「大脑」异步执行长任务,随时可打断。官方评测称 SpokenQA 两个子集居全双工组第一,Daily-Omni 达 78.53%,模型与代码已放出。","7 月 OpenAI 发布 GPT-Live,把「全双工语音 + 异步推理」做成了闭源产品:对话层管听说,推理层管干活。两个月后,开源侧有了对照物——腾讯混元团队 9 月 9 日开源 Gander,一个端到端全模态交互 Agent,模型与代码全部放出,论文同步挂上 [arXiv:2609.08977](https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.08977),进 Hugging Face Daily Papers 日榜第 3,官方仓库 [GitHub](https:\u002F\u002Fgithub.com\u002FOmni-Interaction-Gander\u002FOmni-Interaction-Agent) 发布当天 star 数已过百。\n\n## 双系统:Cerebellum 管「说」,Brain 管「做」\n\nGander 的核心是 Cerebellum-Brain 架构。小脑基于 MiniCPM-o 4.5,负责实时视听感知、全双工对话控制与语音合成;大脑走可插拔 provider 接口,默认接 Codex,异步处理长程推理、工具调用与工作流执行;中间由 Agent 编排运行时衔接,绑定可信对话轮、维护任务状态、管理权限。\n\n交互被压成一秒一个因果单元:[视频][音频][任务上下文] → [控制 token][内容]。控制 token 先于内容预测,把「要不要开口」和「说什么」解耦成 listen \u002F speak \u002F interrupt \u002F tool 四种行为。Talker 与 Thinker 分离部署,按每单元 8 个文本 token 对 50 个 S3 语音 token 的对齐比例流式出声,感知不被语音生成阻塞;长会话保留最近 128 个单元。任务侧有 task_start、task_send(main\u002Ffork)、task_resolve 一套小词汇表,执行代际隔离防止旧任务结果污染新指令。「边说边干活」由此成为原生能力:你可以随时打断,它也能主动汇报进度、追问需求、请求授权。\n\n## 官方评测的数字\n\n按技术报告,评测覆盖对话能力、全模态理解、交互、Agent 四个维度,共 2,052 条基准语句。官方口径的成绩单:SpokenQA 两个子集 75.60 \u002F 59.30,VoiceBench 的 AlpacaEval 3.96 分、SD-QA 46.84%,官方称 SpokenQA 居全双工组两个子集第一、VoiceBench 居第二;Full-Duplex-Bench v3 交互时机 100% 恰当,过早打断率 8.0%;委派场景 45\u002F45 条最终回复正确绑定;加上全双工与 Agent 后训练后 Daily-Omni 仍保持 78.53%。摘要另称内部人类评估确认其口语表达自然,在背景噪声、多方对话、backchannel 场景下保持稳健。\n\n## 该泼的冷水\n\n仓库里数据集徽章写着 coming soon,「模型 + 代码 + 数据」三件套目前只兑现两件;上述评测均为团队自报,第三方复现暂无门路;大脑默认是 Codex,换别家 Agent 要自己接 provider。\n\n真正的增量在于:OpenAI 用闭源产品验证过的「对话层 + 推理层」分工,现在有一条从 MiniCPM-o 4.5 小脑到可插拔大脑的开源实现可以拆开研究——控制 token 协议、运行时的任务生命周期都是白盒。当语音交互从「轮到你说话」变成「随时都能插话」,Agent 的形态也会跟着变:你的工作流,准备好被随时打断了吗?","https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.08977","7437aeb9-930c-4866-a2e9-48003c1a792b",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"6ad31a14-c0da-42df-81fd-564281f768db","agentic-ai",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",{"id":19,"name":20,"slug":20,"description":14,"color":14},"499f4b56-819d-49a3-9609-33e775143b86","multimodal",{"id":22,"name":23,"slug":23,"description":14,"color":14},"b9bd9039-fcdb-41a8-b85b-fc1587def2b9","open-source",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"cab1b70b-1f24-4bb3-8439-d8b666c48446","en","Hunyuan Open-Sources Gander, a Full-Duplex Omni Interaction Agent","Hunyuan open-sources Gander: an omni interaction agent with a full-duplex speech Cerebellum and a pluggable reasoning Brain, built on MiniCPM-o 4.5.","July's GPT-Live turned \"full-duplex speech plus asynchronous reasoning\" into a closed product: a conversation layer for listening and speaking, a reasoning layer for getting work done. Two months later the open-source side has its counterpart — on September 9, Tencent's Hunyuan team released Gander, an end-to-end omni interaction agent, with models and code fully open, the paper on [arXiv:2609.08977](https:\u002F\u002Farxiv.org\u002Fabs\u002F2609.08977), a #3 spot on the Hugging Face Daily Papers board of the day, and a [GitHub](https:\u002F\u002Fgithub.com\u002FOmni-Interaction-Gander\u002FOmni-Interaction-Agent) repo past a hundred stars on release day.\n\n## Two systems: the Cerebellum speaks, the Brain works\n\nGander's core is a Cerebellum-Brain architecture. The Cerebellum, built on MiniCPM-o 4.5, owns real-time audio-visual perception, full-duplex conversation control, and speech synthesis. The Brain sits behind a pluggable provider interface — Codex by default — and asynchronously handles long-horizon reasoning, tool use, and workflow execution. An agent orchestration runtime connects the two, binding trusted turns, tracking task state, and managing permissions.\n\nInteraction is compressed into one-second causal units: [video][audio][task context] -> [control token][content]. The control token is predicted before content, decoupling whether to act from what to produce into four behaviors: listen, speak, interrupt, tool. The Talker deploys detached from the Thinker, streaming audio at a released alignment of eight text tokens to 50 S3 speech tokens per unit, so perception is never blocked by speech generation; long sessions retain up to 128 recent units. The task side exposes a small vocabulary — task_start, task_send (main\u002Ffork), task_resolve — with execution generations fencing stale results after a revised objective. Talk-while-working is thus native: you can interrupt at any time, and it can proactively report progress, ask follow-up questions, or request permission.\n\n## The numbers, vendor-reported\n\nPer the technical report, evaluation covers four dimensions — conversational ability, omni understanding, interactive capability, and agentic intelligence — across 2,052 benchmark utterances. The official scorecard: SpokenQA reaches 75.60 \u002F 59.30 on its two subsets and VoiceBench AlpacaEval scores 3.96 with SD-QA at 46.84%, which the team says place first and second in its full-duplex comparison group respectively; Full-Duplex-Bench v3 shows 100% appropriate turn-taking with an 8.0% premature-interruption rate; delegated scenarios bound 45 of 45 final responses correctly; and Daily-Omni holds at 78.53% after full-duplex and agentic post-training. The abstract adds that internal human evaluations confirm natural, expressive speech that stays robust under background noise, multi-party dialogue, and backchannel communication.\n\n## The cold water\n\nThe repo's dataset badge still reads \"coming soon\" — of the models-plus-code-plus-data triple, only two have shipped. Every number above is self-reported, with no third-party replication path yet. The default Brain is Codex; swapping in another agent means writing your own provider.\n\nThe real increment: the conversation-layer-plus-reasoning-layer split that OpenAI validated inside a closed product now has an open implementation you can take apart — the control-token protocol and the runtime's task lifecycle are all white-box. When voice interaction shifts from \"your turn to speak\" to \"interrupt anytime\", the shape of agents changes with it. Is your workflow ready to be interrupted at any moment?","hunyuan-gander-omni-interaction-agent","2026-09-09T21:07:00Z","2026-09-09T21:07:07.439607Z","2026-09-09T21:07:07.439617Z",true,"agent",204,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"f60a52ae-76c9-4fdf-a1d7-efc41bd50c38","阶跃星辰发布Step 3.7 Flash：面向生产级Agent的开源基座模型升级","stepfun-step-3-7-flash-agent-multimodal","2026-05-29T01:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"dcd8b3e1-a3c7-4614-aba4-9002219ea5f6","LibreDB Studio 0.15 发布:本地 LLM 接管数据库交互","libredb-studio-local-llm-agent","2026-09-15T00:00:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"cd49f913-cde7-4cf3-8d93-24508653180e","腾讯混元开源AuK:1.5B语音模型统一生成与编辑,4步推理快4.5倍","tencent-hunyuan-auk-speech-editing","2026-09-09T09:12:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"44aa8908-1cef-48d6-b224-7de12a8d4afd","NeoHorse-1：让 Agent 执行轨迹进入自我改进回路","neohorse-1-agentic-post-training-rsi","2026-09-09T07:19:25+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"051084dc-3da0-451e-9c6b-a267d5b0e77f","给机器人技能装上门禁:EmbodiedSkills 预检+验证闭环,RoboTwin 50 任务冲到 86.2%","embodiedskills-vla-verify-loop","2026-09-08T17:10:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"9bd3ef9b-d6b9-4793-9156-15ce7d91a66f","微软开源流式说话人 ASR:VibeVoice 7B 实时转录“谁在说什么”","vibevoice-asr-streaming-speaker-attribution","2026-09-07T15:10:00+00:00"]