On-device intelligence for every product为每款产品注入端侧智能

Today we're launching Desert Ant Labs, a European frontier AI lab building opinionated on-device intelligence. We believe the best path to efficient intelligence starts on-device.今天，我们正式推出 Desert Ant Labs，这是一家位于欧洲的前沿 AI 实验室，致力于打造具有鲜明特色的端侧智能。我们坚信，实现高效智能的最佳途径始于端侧。

We're building small, specialized models for audio, vision, and text – each model answers in milliseconds, and costs nothing to run, so you can put intelligence in every product interaction, without being limited by token cost or inference speed. Small enough to run on a five-year-old phone, fast enough to use on every frame or keystroke, and better than the API call you're already paying for.我们正在为音频、视觉和文本任务构建小型化、专业化的模型。每个模型都能在毫秒级时间内做出响应，且运行成本几乎为零。这样一来，你就能将智能融入到每一项产品交互中，而无需受限于 Token 成本或推理速度。这些模型小到足以在五年前的手机上运行，快到足以处理每一帧画面或每一次按键，且性能优于你当前付费使用的 API 调用。

The first 18 models are live today (12 stable and six in beta), accessible via one SDK for Swift, Kotlin, and JavaScript. One model per task, each built to be the fastest way to complete that task on a device:首批 18 个模型已于今日上线（12 个稳定版，6 个测试版），可通过一套适用于 Swift、Kotlin 和 JavaScript 的 SDK 进行调用。每个模型各司其职，旨在成为该任务在设备上运行的最快解决方案：

• Voz: transcribe 10 minutes of audio in two seconds on an iPhone – 4.7x faster than Whisper – with a start and end time on every word.Voz：在 iPhone 上两秒钟即可转录 10 分钟音频，速度比 Whisper 快 4.7 倍，并能标注每个词的起始和结束时间。
• Clear: a 9MB model that can turn a five-minute laptop recording into studio quality audio in one second.Clear：一个 9MB 的模型，能在一秒钟内将笔记本电脑录制的五分钟音频提升至录音室级别音质。
• Redact: mask names, addresses, and card numbers, in real time, in 27 languages, so they never reach your servers.Redact：实时掩码处理姓名、地址和卡号，支持 27 种语言，确保这些敏感数据永远不会发送到你的服务器。
• Tongue: identify 84 languages from three words, with a 2MB model.Tongue：仅需 2MB 的模型，即可通过三个词识别出 84 种语言。

And that's just to name a few. You can find full specs and benchmarks for the other fourteen, on desertant.com/models and Hugging Face. Every model is free up to 100k monthly active devices. No tokens, no logins.这仅仅是其中几个例子。你可以在 desertant.com/models 和 Hugging Face 上查看其余 14 个模型的完整规格和基准测试。每个模型在每月 10 万台活跃设备内均可免费使用。无需 Token，无需登录。

We're building this in Europe, where "on-device" is the sovereign default. The data never leaves your customer's hands, the feature never depends on someone else's cloud, and what's never been uploaded can never be compelled.我们在欧洲构建这些技术，在这里，“端侧”是数据主权的首选默认设置。数据永远不会离开客户手中，功能无需依赖他人的云服务，从未上传的数据也就永远无法被强制调取。

How we got here我们的历程

For five years we've been building our video app, Detail, with an on-device first approach. But when we introduced features like Auto Edit to create short clips, or audio enhancement for podcasts, we had to fall back to cloud APIs. And as the popularity of Detail grew, so did our infrastructure bills.过去五年里，我们一直在开发视频应用 Detail，并坚持“端侧优先”的理念。但当我们引入诸如自动剪辑（Auto Edit）或播客音频增强等功能时，不得不退而求其次使用云端 API。随着 Detail 的流行，我们的基础设施账单也随之水涨船高。

Every few months I'd hunt for useful on-device models. I'd surf Hugging Face for a model that could find filler words or clean up a recording. And, every June, we'd get great new tools to build with but the industry wasn't moving fast enough. The foundation was there: the chips, Core ML, the research. What was missing was everything between that foundation and actually implementing a feature in your app: a model you could drop in and ship with a few lines of code.每隔几个月，我就会寻找有用的端侧模型。我会浏览 Hugging Face，寻找能去除语气词或优化录音的模型。每年六月，我们都能获得强大的新工具，但行业的发展速度仍不够快。基础条件已经具备：芯片、Core ML 以及相关研究。缺失的是连接这些基础与实际功能实现之间的桥梁——即一个你可以直接嵌入并用几行代码即可部署的模型。

So, we trained the models ourselves. It turns out training a model is a product design challenge, and product is what we know. We designed models and local inference that beat cloud services on speed, quality, and cost, and outperform other local and cloud models on the task itself, at a fraction of their size.于是，我们亲自训练了这些模型。事实证明，训练模型本身就是一项产品设计挑战，而产品正是我们的强项。我们设计了模型和本地推理方案，在速度、质量和成本上击败了云服务，并在特定任务上超越了其他本地和云端模型，而体积仅为其一小部分。

We replaced Dolby for better, faster audio enhancement with Clear, and made our on-device transcriptions 5x faster with Voz. We also replaced Claude Sonnet with Clips, our 284MB model that turns a 10-minute video into a dozen clips in 5 seconds – 10x faster and using 470x less energy than Sonnet, with the same quality.我们用 Clear 取代了 Dolby，实现了更好、更快的音频增强；利用 Voz 将端侧转录速度提升了 5 倍。我们还用 Clips 取代了 Claude Sonnet——这是一个 284MB 的模型，能在 5 秒内将 10 分钟的视频剪辑成十几段短片，速度比 Sonnet 快 10 倍，能耗降低 470 倍，且质量保持一致。

Detail 6, which will launch with iOS 27, replaces all of our cloud APIs with our own models, running entirely on the device.Detail 6 将随 iOS 27 发布，它将用我们自己的模型替换掉所有云端 API，完全在设备上运行。

We've all spent the past few years building with LLMs as if they were just another API. And, amid the hype around generalist frontier brains, we almost forgot they're not the only option.过去几年，我们都习惯将大语言模型（LLM）仅仅当作又一个 API 来使用。在围绕通用前沿 AI 的炒作中，我们几乎忘记了它们并非唯一的选择。

Every developer I talk to has a wishlist of on-device models they'd build if cost wasn't a factor, or a feature they're bleeding tokens on that they'd happily swap for a local model. A call that runs the same way a hundred thousand times a day: cleaning a recording, tagging a photo, pulling a date out of a sentence, catching a name before the text hits your servers. None of these needs a frontier model.我接触过的每一位开发者都有一个心愿清单：如果成本不是问题，他们想构建哪些端侧模型；或者他们正在为哪些功能支付高昂的 Token 费用，而这些功能本可以轻松替换为本地模型。比如：清理录音、标记照片、从句子中提取日期、在文本发送到服务器前过滤姓名。这些任务根本不需要前沿大模型。

NVIDIA's own researchers pulled apart three agent systems and estimated that 40 to 70% of their calls to a large model could go to a small, specialized one instead.NVIDIA 的研究人员拆解了三个智能体系统，评估发现其中 40% 到 70% 的大模型调用本可以由小型、专业化的模型来完成。

The compute is already paid for算力成本已被支付

The industry will spend about $450 billion on data centers this year. Meanwhile, the world ships more than a billion phones, tablets, and laptops with increasingly capable chips, perfectly suited to these kinds of tasks. There's more compute available in people's hands than in every AI data center on earth.今年，整个行业将在数据中心上投入约 4500 亿美元。与此同时，全球每年出货超过十亿部手机、平板电脑和笔记本电脑，它们配备的芯片性能日益强大，完全能够胜任这些任务。人们手中拥有的算力总量，远超地球上所有 AI 数据中心的总和。

We have an unfair advantage with free inference. No per-call cost, so a feature runs on every message instead of the ones you can afford to check. No round-trip, and your customer's data never leaves the device. When inference costs nothing, the way we build products changes entirely.我们在免费推理方面拥有不公平的竞争优势。没有单次调用成本，因此功能可以运行在每一条消息上，而不是仅限于你负担得起检查的那部分。没有往返延迟，客户数据也永远不会离开设备。当推理成本为零时，我们构建产品的方式将彻底改变。

Little brains in every product为每款产品注入智慧大脑

To build with local models, the developer experience has to get a lot better. You need models you can use commercially, that beat the alternatives on your task in speed and quality, that you can drop into your app with a few lines of code, and are easy to discover.要使用本地模型进行开发，开发者体验必须大幅提升。你需要能够商用的模型，在速度和质量上胜过替代方案，只需几行代码即可嵌入应用，且易于发现。

Think of the first hundred models as the cerebellum, the little brain. The little brain handles the always-on work – balance, timing, the skills you never think about, so the rest of the brain is free to think. That's what we're building first: fast, specialized models for the work that runs all day, on the device, for free.将首批百个模型想象成“小脑”。小脑处理那些常驻的后台工作——平衡、时序、那些你无需思考的技能，从而让大脑的其他部分腾出空间去进行深度思考。这正是我们首先要构建的：快速、专业化的模型，用于处理那些全天候运行、在设备端免费执行的任务。

Then comes the cortex, the layer that decides which model answers. A small local model first, a bigger one when the job requires it, and the cloud only when the work has to leave the device. As open research advances and device silicon becomes more capable, the local models grow, and we'll train larger ones ourselves. Frontier intelligence, built from the small end up.接下来是“大脑皮层”，即决定由哪个模型来回答的层级。首先使用小型本地模型，当任务需要时使用更大的模型，仅在不得不离开设备时才调用云端。随着开源研究的进步和设备芯片能力的提升，本地模型将不断壮大，我们也会亲自训练更大的模型。从小型模型开始，自下而上构建前沿智能。

Cloud labs ship neutral models because per-token pricing needs a neutral model. Every Desert Ant model ships with a default we choose, and the levers you need to change that default. We optimize the model and the runtime together: on an iPhone, Clear and Voz run on the Neural Engine, and in the browser, Clear's same weights run through WebAssembly.云端实验室提供中性模型，因为按 Token 定价模式需要中性模型。每个 Desert Ant 模型都带有我们选择的默认设置，并提供你修改这些设置所需的控制杠杆。我们同时优化模型和运行时：在 iPhone 上，Clear 和 Voz 运行在 Neural Engine 上；在浏览器中，Clear 使用相同的权重通过 WebAssembly 运行。

The SDKSDK

Ready to get started? You can implement Desert Ant models in your app with our native Swift, Kotlin, and JavaScript SDK, available on GitHub.准备好开始了吗？你可以通过 GitHub 上提供的原生 Swift、Kotlin 和 JavaScript SDK 将 Desert Ant 模型集成到你的应用中。

Our docs are written for developers and agents and you can try the models on your Mac with the CLI, or in your browser on Hugging Face.我们的文档专为开发者和智能体设计，你可以通过 CLI 在 Mac 上试用这些模型，或者在 Hugging Face 的浏览器中体验。

Building something cool with our models, or want to build them with us? Get in touch.正在用我们的模型构建酷炫的应用，或者想与我们合作开发？欢迎联系我们。