小艺Claw,终于不是旗舰专属了
今天有个事儿挺有意思——华为宣布小艺Claw全机型开放了。之前小艺Claw只在Mate系列和P系列旗舰机上才能用,说白了就是"买得起旗舰才配用"。现在HarmonyOS 5.0及以上的设备都能用,从千元机到平板、手表、车机,全给你安排上。
这个动作背后,华为的逻辑很清楚:AI助手不是奢侈品,是基础设施。你看小艺的进化路线就能看出来。2023年还是基础版语音助手,2024年上了大模型,2025年深度整合鸿蒙,到2026年直接全机型铺开。每一步都在往"让更多人用上"这个方向走。这次开放的时间点很微妙。鸿蒙生态刚突破一定规模,这时候把AI能力下沉到全系设备,等于给所有鸿蒙用户发了一张"AI入场券"。你买了华为设备,AI助手就是标配,不用额外付费,不用等OTA推送。我觉得这招挺聪明的——用AI能力绑住用户,比用硬件参数绑用户管用多了。
这次开放,具体给了什么
小艺Claw这次开放的不是阉割版,是完整版。
六个核心能力全给你:多轮对话能记住上下文,跨应用操作靠鸿蒙分布式能力打通,文档处理可以智能摘要,图像识别拍照即搜,日程管理智能提醒,设备控制覆盖全屋智能。说个实际场景你就懂了。你对着手机说"帮我把昨天开会那张照片里的文字提取出来,发到工作群里",小艺Claw能跨相册、OCR、社交应用三个环节一气呵成。这在以前得手动切好几个App。跨应用操作这个能力,是目前其他手机AI助手比较难做到的。因为它是鸿蒙分布式架构的原生能力,不是外挂一个API调用。系统级的权限和调度,第三方App做不到这个深度。当然,实际体验还得看具体机型的性能。千元机跑大模型肯定不如旗舰机流畅,但华为做了端云协同——复杂任务上云端,简单任务本地跑,尽量让低端机也能用。
哪些设备能用上
覆盖面比你想的大。
手机端,Mate系列和P系列全系支持,nova系列部分机型也在列。平板端MatePad全系覆盖。手表端Watch GT系列能用。车机端问界全系安排上了。智慧屏Vision系列也没落下。五个品类全打通,这在手机厂商里不多见。其他家要么只做手机端,要么设备多但AI能力参差不齐——同一家的AI助手在手机和音箱上体验差距挺大。华为这次是一套AI能力打穿所有设备,手机上调好的模型,手表上也能用,车机上也能用。这个"一套能力全设备"的打法,核心靠的是鸿蒙的分布式架构。同一个AI引擎跑在手机、手表、车机上,只是根据设备算力做适配,底层逻辑是通的。对用户来说意味着什么?你在手机上跟小艺Claw聊到一半,上车后车机自动接上,到家智慧屏继续。跨设备无缝衔接,这是鸿蒙生态真正的护城河。
跟其他AI助手比,差在哪好在哪
放到整个手机AI助手的牌桌上,小艺Claw的位置挺有意思。
Siri的优势是iOS生态和隐私保护,但跨应用操作能力一直偏弱,基本停留在"打开某个App"的水平。小爱同学的优势是米家设备多、覆盖广,但AI能力在不同设备上割裂感比较明显——手机上的小爱和音箱上的小爱,像是两个产品。小艺Claw的优势在哪?一个是鸿蒙的深度整合,跨应用、跨设备操作是原生的。另一个是盘古大模型做底层,不是套壳第三方模型。但短板也有。鸿蒙生态的设备总量跟iOS和安卓比还有差距,第三方App的适配也不如另外两家完善。小艺Claw再好用,如果你的常用App还没适配鸿蒙,体验就会打折扣。说白了,这是一场"深度 vs 广度"的竞争。华为选的是深度——在自家生态里做到极致。另外两家选的是广度——覆盖更多设备和场景。
鸿蒙AI的野心,不止一个手机助手
小艺Claw全机型开放,只是鸿蒙AI战略的一个切面。
往上看,小艺Claw是用户交互层,你摸得到、用得着。往下走,盘古大模型是AI能力层,提供语言理解、图像识别、多模态能力。再往下,鸿蒙内核是系统调度层,负责把AI任务分配到最合适的设备上跑。最底层是全场景硬件,手机、平板、手表、车机、智慧屏。这四层架构的核心逻辑是:AI能力不是某个App的功能,而是操作系统的原生能力。华为想做的事很清楚——让AI像WiFi一样,成为设备的默认能力。你买设备,AI就在那里,不用装App,不用注册账号,不用付费订阅。这个思路跟其他厂商不太一样。有的厂商把AI做成一个App,有的做成一个API,华为是直接把AI做到系统内核里。当然,这条路能不能走通,最终看两点:一是鸿蒙生态能不能持续扩大,二是盘古大模型的能力能不能持续领先。但至少方向是对的。AI助手从"旗舰专属"变成"全系标配",这个趋势不可逆。华为只是先走了一步。
📌 关注「今朝看AI」,每天带你追AI最新动态。
小艺Claw, no longer exclusive to flagships
Today, Huawei announced the full-device availability of Xiaoyi Claw. Previously, Xiaoyi Claw was only available on flagship models like the Mate and P series, which were essentially “only for those who can afford the flagship.” Now, with HarmonyOS 5.0 and above, devices from budget phones to tablets, watches, and car systems can all use it, fully equipped. Behind this move, Huawei’s logic is clear: AI assistants are not luxury goods — they’re infrastructure. You can see this in Xiaoyi’s evolution: from a basic voice assistant in 2023, to integrating large models in 2024, deep integration with HarmonyOS in 2025, and full-device coverage in 2026. Each step moves toward “making it available to more people.” The timing of this release is subtle — just as the Harmony ecosystem reaches a certain scale, lowering AI capabilities to all devices gives every Harmony user an “AI entry ticket.” Bought a Huawei device? AI assistant is standard, no extra fees, no OTA waiting. This is a smart strategy — tying users with AI is more effective than with hardware specs.
What Exactly Does This Release Include?
The release is not a cut-down version; it’s the full monty. Six core capabilities are included: multi-turn dialogue with context memory, cross-app operations via distributed HarmonyOS architecture, document analysis (summarizing text from images), image recognition (search by photo), calendar management (smart reminders), and device control (home automation). For example, you can ask your phone to extract text from a meeting photo taken yesterday and send it to a work group — 小艺Claw handles the entire chain: gallery access, OCR, messaging app integration. The cross-app operation feature is particularly notable because it's built directly into the operating system, not a third-party plugin. While lower-end devices may have slower processing, Huawei compensates with cloud-edge collaboration — moving heavy tasks to the cloud while keeping light ones local.
Which Devices Are Supported?
The coverage is more expansive than you might think. On mobile, the entire Mate and P series are supported, along with select nova models. On tablets, the MatePad series is fully covered. For wearables, the Watch GT series gets the feature. In cars, the entire AITO (问界) lineup is equipped, and even smart displays (Vision series) are included. Unifying five product categories is rare in the industry; other companies often limit AI to one or two categories or fail to synchronize features across them. The key here is consistency — the same AI engine runs across all devices, adapting to their respective hardware capacities. So if you start a conversation on your phone, hop in the car, and continue it, the car knows exactly where you left off.
How Does It Compare to Other AI Assistants?
In the landscape of smartphone AI assistants, 小艺Claw occupies a unique niche. Siri excels in iOS integration and privacy but lacks robust cross-app automation. Xiaomi's 小爱同学 has broad hardware coverage but suffers from fragmented experiences — a different小爱 on your phone vs. your TV. 小艺Claw's strength lies in its tight integration with HarmonyOS — native support for crossing apps and devices. It also leverages Huawei's own Pan Gu model, ensuring consistency. However, its limitation is that the Harmony ecosystem is still smaller than iOS or Android, meaning fewer third-party apps are optimized for the experience. So while it has depth, it doesn't yet have the breadth of its competitors — a trade-off between "doing one thing well" within a limited ecosystem versus broad, somewhat fragmented coverage.
