Anker has introduced the soundcore Liberty 5 Pro and Liberty 5 Pro Max, a pair of true wireless earbuds built around the company’s first custom neural-net chip, called Thus. The chip performs real-time noise cancellation directly on the earbud hardware, with no need to stream audio to a cloud server. A Guinness World Records verification dated 23 April 2026 gave the Liberty 5 Pro a G-MOS speech-quality score of 3.68, the highest objective result ever recorded for true wireless earbuds.
Why a dedicated AI chip in earbuds matters right now
Most noise-canceling earbuds rely on conventional digital signal processors that run fixed filter algorithms. Those designs work well enough in steady-state environments like airplane cabins, but they struggle with unpredictable, rapidly shifting noise such as street traffic mixed with wind and crowd chatter. Anker’s answer is to move a trained neural network onto a purpose-built chip small enough to fit inside an earbud housing. Because the model runs locally, the system can adapt its cancellation profile in milliseconds without the latency penalty of a round trip to a remote server, according to soundcore documentation.
The underlying architecture, compute-in-memory, tackles a specific bottleneck. In traditional processors, energy is consumed every time data shuttles between storage and the arithmetic unit. Compute-in-memory chips perform calculations directly inside the memory array, cutting that shuttle cost. A peer-reviewed study in Nature showed how resistive memory arrays can store neural-network weights in cells that double as computing elements. Anker says the Thus chip applies this principle to audio workloads, keeping power draw low enough for a battery-constrained earbud.
That power claim raises a testable question: will compute-in-memory designs show at least 30 percent lower average power draw during continuous noise cancellation compared with DSP-only earbuds when measured across identical usage profiles? No independent lab has published such a head-to-head comparison for the Liberty 5 Pro. The hypothesis is technically plausible, given that the Nature research documents significant energy savings from eliminating data movement, but confirming the real-world margin will require controlled third-party benchmarks that have not yet appeared.
Speech-quality evidence and the Guinness G-MOS record
The strongest independent data point so far is the Guinness World Records entry. The organization verified on 23 April 2026 that the ANKER soundcore Liberty 5 Pro achieved a G-MOS score of 3.68, the highest speech quality score for TWS earbuds in an objective test. G-MOS, short for Generalized Mean Opinion Score, is a standardized metric that quantifies how clearly a listener’s voice reaches the far end of a phone call under noisy conditions. A higher score means the microphone system and its processing pipeline do a better job of isolating speech from ambient sound.
Anker ties that result directly to the Thus chip’s on-device neural network, which filters environmental noise from the microphone feed before it reaches the Bluetooth codec. The company’s product announcement frames the chip as the reason the Liberty 5 Pro can maintain call clarity in conditions that would overwhelm a conventional beamforming microphone array. The Liberty 5 Pro Max shares the same Thus silicon, though the Guinness certification names only the Liberty 5 Pro model.
What makes the G-MOS figure meaningful is its objectivity. Unlike subjective listening panels, the test uses an automated measurement rig that applies a defined noise profile and scores the output against a reference signal. That removes brand bias from the result. Still, a single metric does not capture every dimension of audio performance. Latency, frequency response, and battery life under sustained cancellation are separate variables that the G-MOS protocol does not address.
Open questions about Thus and real-world battery impact
Several gaps remain between the marketing narrative and verifiable field performance. First, Anker has not released raw test logs or detailed latency measurements showing how quickly the neural network adapts to changing noise environments. The Guinness record confirms speech quality at one measurement point, but it does not describe how the system behaves over hours of continuous use or across varied acoustic scenes such as a busy cafe, a subway platform, or a windy outdoor path.
Second, the peer-reviewed Nature paper on resistive RAM compute-in-memory describes the architecture’s general energy advantages, but it contains no device-specific measurements for the Thus chip or the Liberty 5 Pro hardware. The gap between laboratory-demonstrated resistive RAM arrays and a shipping consumer product is significant. Without published power-draw curves from an independent testing house, the 30-percent efficiency hypothesis remains an educated guess rather than a confirmed fact.
Third, no independent audio engineers or third-party labs have published direct comparisons between Thus and competing on-device AI processors from companies like Qualcomm or Apple. Both of those firms have integrated machine-learning accelerators into their mobile platforms, and several earbuds already lean on those cores for adaptive ANC and voice processing. Until controlled benchmarks pit Liberty 5 Pro against rivals under identical noise profiles, claims of categorical leadership will be hard to assess.
Battery life is another unresolved area. Anker advertises multi-hour playback with ANC enabled, but marketing figures typically assume moderate volume and intermittent use. The compute-in-memory design should, in theory, reduce the energy cost of running a neural network continuously. In practice, total battery life also depends on radio efficiency, driver power, and how aggressively the firmware throttles the AI workload when the acoustic scene is stable. Long-duration stress tests in real cities and transit systems will be needed to reveal whether Thus meaningfully extends runtime compared with traditional DSP-based designs.
How Anker is positioning Liberty 5 Pro and 5 Pro Max
In its launch materials, Anker emphasizes that the Liberty 5 Pro line is meant to showcase earbuds that “think,” highlighting on-device intelligence rather than just incremental sound tuning. The company describes the Thus chip as a neural processing core that can run multiple models, including adaptive noise cancellation, wind suppression, and context-aware transparency modes. According to Anker’s announcement, the 5 Pro Max builds on the base model with additional microphone arrays and refinements to spatial audio, while sharing the same underlying silicon.
This positioning reflects a broader shift in the audio market. As smartphones and laptops gain dedicated AI accelerators, accessory makers are looking for ways to differentiate beyond codec support and driver materials. By putting a neural chip directly in the earbud, Anker is betting that users will value consistently clear calls and adaptive sound profiles as much as, or more than, raw audio fidelity. The Guinness G-MOS record gives that story a concrete benchmark, even if it does not answer every question about music playback quality or long-term durability.
At the same time, the Liberty 5 Pro and 5 Pro Max enter a crowded field where premium earbuds already promise “smart” features such as automatic conversation detection and personalized ANC. Some of those functions run primarily on the paired phone, while others use lighter-weight algorithms on the earbuds themselves. Thus represents a more aggressive step toward self-contained processing. If third-party testing confirms that this approach can deliver better speech intelligibility without sacrificing battery life, it could nudge competitors toward similar compute-in-memory architectures.
What to watch for in early reviews and lab tests
For prospective buyers and industry watchers, several data points will matter more than launch-day claims. Independent measurements of call quality in varied noise environments will show whether the Guinness-certified performance generalizes beyond a controlled test rig. Side-by-side battery benchmarks against leading ANC earbuds will help quantify any efficiency gains from compute-in-memory. Objective audio tests, including frequency response and distortion, will reveal whether the focus on AI has come at the expense of traditional sound engineering.
Equally important will be transparency around firmware updates. Neural models can be retrained and redeployed, potentially improving performance over time or tailoring behavior to new use cases. Anker’s long-term support strategy for Thus-powered products will determine whether Liberty 5 Pro and 5 Pro Max remain competitive as rival earbuds adopt their own on-device AI hardware. For now, the combination of a record-setting G-MOS score, a custom compute-in-memory chip, and a clear marketing push toward “thinking” earbuds makes this launch one of the more technically ambitious entries in the true wireless space, even as key questions about power, latency, and comparative performance remain open.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
