Voice calls placed over 4G networks deliver better quality than calls carried on 5G in most locations, according to a Light Reading analysis of call performance data, a finding that undercuts the assumption that each network generation improves on the last for every service.

The result matters commercially. Operators have spent years and billions promoting 5G as a uniform upgrade, yet the voice layer, the service subscribers use most minutes on, still runs better on the older generation across much of the footprint. Call quality rankings of this kind feed directly into operator marketing, network awards and churn arguments.

Why would 4G beat 5G on voice?

The answer lies in maturity, not physics. Voice over LTE, the 4G voice standard known as VoLTE, has been deployed at scale for close to a decade. Operators have tuned it, device support is near-universal, and fallback mechanisms are well understood. Voice over New Radio, the 5G equivalent known as VoNR, is far younger. Few operators run VoNR as the default voice path, coverage is patchier, and handovers between generations still introduce friction.

Where a 5G phone cannot find a reliable VoNR bearer, the network drops the call back to 4G or even to legacy circuit-switched fallback. That transition, not the radio itself, is where quality suffers. The handset also plays a part: 5G devices juggling non-standalone architectures can handle voice differently from the 4G-only fleet.

What does this change for operators?

Practically, it tempers the 5G pitch. An operator can claim national 5G coverage and still post weaker voice scores than its 4G baseline, because the measurement follows the call, not the logo on the coverage map. Marketing departments that lean on generational labels should note the gap; subscribers judge networks by whether calls drop, not by which bearer carries them.

The finding also sets a benchmark. As operators push standalone 5G and switch on VoNR more widely, the natural test of progress is simple: when does 5G voice stop losing to 4G? Operators that close the gap first gain a measurable claim in network-quality comparisons, the kind that industry awards and enterprise contracts cite.

Standards context in brief

3GPP defined VoNR as the native voice path for standalone 5G, with VoLTE and circuit-switched fallback as interim bridges in non-standalone deployments. Most commercial 5G networks launched in non-standalone mode, anchored to a 4G core, which means most 5G-branded calls today ride on 4G voice machinery anyway. Standalone rollouts, where VoNR actually applies, remain a minority of deployments globally.

That standards history explains the paradox. A "5G call" on a non-standalone network is often a 4G call with a 5G data session attached. Only where standalone 5G carries native VoNR does the comparison become a true generation-versus-generation test, and there the newer standard is still climbing its learning curve.

Subscribers will not notice, for now

For consumers, the practical effect is limited. Calls on both generations work, and the quality difference shows up in measurement rankings rather than in everyday complaints. The divergence matters more to operators, benchmarking firms and device makers than to the average user placing a call.

The direction of travel, however, is clear. Each VoNR deployment adds tuning data, device support widens with every chipset generation, and standalone cores expand. Light Reading's finding captures a moment in that transition, not a permanent verdict: 4G voice holds the lead in most places today, and the industry's own rollout timelines for standalone 5G will decide how quickly that lead changes hands.