A new report from Ookla®, published October 4, 2026, is upending a core assumption of the 5G era: that newer radio technology automatically means better phone calls. Ookla’s study, titled “The 5G Voice Paradox: Why VoNR Isn’t Always Better (Yet),” found that in most markets tested, the decade-old VoLTE standard running on 4G still connects calls faster and more consistently than VoNR, the voice technology built for standalone 5G. The finding runs counter to years of carrier marketing suggesting 5G improves every part of the mobile experience, including the basic act of placing a phone call.
The report lands at a moment when carriers worldwide are racing to activate 5G Standalone networks and retire older infrastructure. Ookla’s data suggests that race may be getting ahead of the technology’s actual performance on something as fundamental as voice. For an industry that spent the better part of a decade promising 5G would fix everything LTE could not, a report concluding that calls over 4G still beat calls over 5G in most places is the kind of finding that reorders expectations.
What Ookla’s October Report Actually Found
Ookla’s key takeaway, published alongside the October 4 report, states plainly that VoLTE remains the gold standard for consistency and speed in mature deployments. The company’s flagship data point comes from South Korea: operators in Seoul recorded a 1.06-second median call setup time on VoLTE, a result Ookla says was faster than any VoNR deployment studied anywhere in the world. More than 90% of the calls measured in that Seoul VoLTE sample met Ookla’s two-second call-setup benchmark, the threshold the firm treats as the standard for a responsive voice experience.
That is a striking result, because VoNR, short for Voice over New Radio, is supposed to be the more advanced technology. It runs natively on 5G Standalone networks rather than relying on a 4G core, and in theory it should produce lower latency and faster signaling than VoLTE, or Voice over LTE, which has been the default method for carrying voice calls over 4G networks since the early 2010s. Ookla’s testing suggests that theory and practice have not yet lined up, at least not everywhere.
According to Ookla’s and Omdia’s joint 2026 report on 5G Standalone and 5G Advanced deployments, “In Ookla’s global testing, VoLTE had a median call setup time of 1.87 seconds, compared with 1.96 seconds for VoNR.” That is a global, aggregated figure spanning multiple operators and markets, and the gap between the two technologies is narrow in absolute terms. But the direction of that gap, with the older technology coming out ahead, is what makes the finding newsworthy. You can read the underlying data from Ookla’s RootMetrics research program, which has tracked US network performance using similar methodology.
VoLTE and VoNR, Explained Simply
The distinction matters because most phone users have no idea which voice technology their call is actually running on. VoLTE carries voice traffic as data packets over a 4G LTE connection using an IP Multimedia Subsystem, rather than the older circuit-switched method that defined 2G and 3G calling. It has been the default voice technology on most modern smartphones for roughly a decade, which means carriers have had years to tune, optimize, and debug it.
VoNR performs the same basic job, moving voice as IP data rather than circuit-switched traffic, but it does so entirely within a 5G Standalone network, without dropping down to a 4G core at any point in the call setup process. On paper, that should be an advantage: 5G Standalone networks offer lower theoretical latency and more granular quality-of-service controls than 4G. In practice, VoNR depends on having continuous, well-optimized 5G Standalone coverage along the entire call path. Where that coverage is incomplete, calls may fall back to 4G anyway, adding delay rather than avoiding it.
That dependency is central to understanding why Ookla’s results cut against the simple “5G is always faster” narrative. VoNR’s performance ceiling is high, but reaching it requires a level of network maturity that most operators, even ones with extensive 5G marketing campaigns, have not yet built out everywhere they claim coverage.
South Korea Sets the Global Benchmark
Seoul’s 1.06-second VoLTE median is the headline number in Ookla’s report, but the country-level detail behind it is just as telling. South Korea has some of the densest, most mature LTE infrastructure in the world, built up over more than a decade of continuous investment by SK Telecom, KT, and LG U+. That maturity appears to show up directly in the call-setup numbers.
Separate RootMetrics testing from South Korea in the second half of 2025, run under controlled conditions, found LG U+ posting the fastest voice call setup time among Korean carriers at 0.692 seconds, connecting calls nearly twice as fast as KT’s 1.309 seconds in the same test cycle. A second measurement round from the same period again put LG U+ ahead, at 0.714 seconds, with SK Telecom following at 0.969 seconds and KT trailing at 1.404 seconds. Those figures are specific to controlled test conditions rather than the broader Ookla global dataset, but they point in the same direction: in South Korea’s case, the fastest, most reliable calling experience is still happening over mature 4G infrastructure, not the newer 5G Standalone networks the same carriers are simultaneously building out.
KT’s own comparison between its two voice technologies, cited in reporting on the Ookla data, found a VoLTE median call setup time of 1.21 seconds against 1.47 seconds for VoNR, meaning the carrier’s 4G voice product outperformed its newer 5G voice product by roughly 0.26 seconds. That is not a small gap in an experience most users judge in fractions of a second; call setup delay is one of the few network performance metrics ordinary people can actually feel.
Where VoNR Actually Wins: Singapore and Malaysia
Ookla’s report is careful not to frame VoNR as universally worse, and the data backs that nuance up. Across the eight operators Ookla studied with statistically meaningful results, six connected calls faster using VoNR than VoLTE, a detail that complicates any simple “4G always wins” headline even as the global median favors VoLTE.
Singapore is the clearest example of VoNR delivering on its promise. Singtel recorded a median VoNR call setup time of 1.07 seconds, compared with 1.22 seconds for VoLTE on the same network, a result in which the newer technology genuinely outperformed the older one. More than 90% of Singtel’s VoNR samples met Ookla’s two-second call setup target, matching the kind of consistency usually associated with mature VoLTE deployments. Malaysia’s Yes 5G network produced similar results, with reporting on the Ookla data noting that the operator “gets VoNR right” even as the broader regional and global picture remains mixed.
Those results matter because they show VoNR’s underperformance elsewhere is not an inherent limitation of the technology. It is a deployment and optimization problem. Carriers that have invested heavily in dense, well-tuned 5G Standalone coverage, like Singtel and Yes 5G, are already beating VoLTE on the same metric where other operators are still lagging behind their own 4G networks.
The StarHub Outlier: When 5G Voice Falls Behind
Not every carrier managed Singtel’s result, even within the same market. StarHub’s VoNR median call setup time came in at 1.82 seconds, 32% slower than its own VoLTE performance on the same network. Just 57% of StarHub’s VoNR call samples met the two-second target, compared with nearly 80% of its VoLTE samples clearing that bar. That is a meaningful consistency gap between two carriers operating in the same small, dense, 5G-saturated country, which suggests the difference comes down to network engineering choices rather than geography or device availability.
The StarHub figures are a useful corrective to any assumption that a carrier’s 5G marketing claims translate directly into a better calling experience. Two operators serving the same city, with broadly comparable spectrum holdings and infrastructure budgets, can land on opposite sides of the VoLTE-versus-VoNR performance gap depending entirely on how carefully they have tuned their standalone 5G voice path.
Why Enterprise Voice Platforms Still Default to 4G
The pattern Ookla documented at the carrier level is already shaping product decisions elsewhere in the industry. Aircall, which builds cloud calling software used by businesses, currently tells its customers that 4G LTE offers more consistent call quality than 5G for real-time voice applications, and that most networks still rely on 4G for critical voice functions even as 5G data coverage expands. The company’s support documentation, titled “Why 4G LTE provides the most reliable mobile VoIP experience,” points to limited availability of true 5G Standalone voice as the core reason, noting that most regions still route voice through VoLTE rather than native VoNR, even on networks that otherwise advertise comprehensive 5G coverage.
Aircall also recommends 4G LTE-only operation in certain mobile VoIP scenarios specifically to avoid the added complexity of a device switching between 5G and 4G mid-call, which it identifies as a potential source of quality variation. The company frames 4G’s broader geographic coverage, particularly outside dense urban cores, as a major reason LTE voice remains more dependable overall, a point that lines up with Ookla’s operator-level findings showing the carriers with the most mature, widest LTE footprints also posting the fastest, most consistent call setup numbers.
A Brief History of How Mobile Calls Got This Complicated
Phone calls were circuit-switched for decades before any of this became an issue. 2G and 3G networks reserved a dedicated channel for the duration of a call, a model that was simple and predictable but inefficient for carrying anything beyond voice. The shift to all-IP networks starting with 4G LTE forced carriers to invent a new way to carry voice as data, which is how VoLTE came into existence in the early 2010s. Early VoLTE deployments were rough, plagued by dropped calls and inconsistent audio, but a decade of iteration turned it into the reliable baseline Ookla’s October report now credits as the “gold standard.”
VoNR is going through the same growing pains VoLTE went through roughly ten years ago, except this time the comparison point is a mature, heavily optimized predecessor rather than clunky circuit-switched calling. That is a much higher bar to clear. VoLTE did not have to beat anything nearly as refined as itself when it launched; VoNR does, which helps explain why Ookla’s October 2026 report reads less like a condemnation of 5G voice and more like a snapshot of a technology still catching up to the standard it was designed to replace.
Market Impact: What This Means for Carrier Marketing
Carriers worldwide have spent billions on 5G Standalone rollouts and matching advertising campaigns built around the promise that 5G improves everything, voice included. Ookla’s report complicates that pitch in markets where VoNR is still underperforming VoLTE, because call quality and setup speed are among the only network metrics an average subscriber can actually perceive in daily use. Download speed benchmarks are mostly invisible to someone making a phone call; a slow or dropped call is not.
For operators like Singtel and Yes 5G that have already closed the gap, the Ookla data is a genuine marketing asset, evidence that their 5G investment is paying off in a way customers will notice. For carriers still posting KT-style results, where VoNR trails VoLTE, the smarter short-term move may be to keep routing voice through VoLTE by default while continuing to tune standalone 5G voice quietly in the background, rather than pushing VoNR to subscribers before it is actually ready to outperform the service it is meant to replace.
Competitive Comparison: VoLTE, VoNR, VoWiFi, and CSFB
Voice over 4G and 5G are not the only ways calls get carried on modern networks, and it is worth placing Ookla’s findings in that broader context. Circuit-switched fallback, or CSFB, is the oldest fallback method still active on some networks, dropping a device from LTE data back to a 2G or 3G voice channel for the duration of a call; industry guides describe CSFB as typically adding five to eight seconds to call setup time, which is why carriers have worked for years to move subscribers off it entirely. VoWiFi, or voice over Wi-Fi, routes calls through a broadband internet connection instead of cellular radio and is commonly used as a coverage extender in buildings with weak signal.
Against that backdrop, VoLTE’s roughly one-to-two-second median setup times and VoNR’s similar range both represent a massive improvement over CSFB-era calling. The real competitive story in 2026 is not VoLTE or VoNR against the old circuit-switched methods, which both comfortably beat; it is the two IP-based standards competing against each other for the final fraction of a second, a contest where maturity and optimization currently matter more than raw radio generation.
The Technical Reasons VoNR Still Struggles in Most Markets
VoNR’s underperformance in Ookla’s global median is not a flaw in the underlying 5G New Radio architecture. It stems from a handful of deployment realities that vary by operator. Full 5G Standalone coverage, meaning a 5G core network paired with 5G radio access rather than a 5G radio layer bolted onto a 4G core, is still incomplete in many regions, which forces devices to fall back to LTE mid-call and erases any latency advantage VoNR might have delivered. Device and chipset support for VoNR also lags behind VoLTE, since every modern phone supports the latter but VoNR compatibility depends on specific hardware and software combinations that are still rolling out.
Mobility handling, meaning how gracefully a call survives a subscriber moving between 5G cells or between 5G and 4G coverage, is another area where VoLTE’s decade of field tuning gives it an edge that VoNR has not yet matched everywhere. None of these are permanent limitations. They are the same category of problems VoLTE itself worked through in its first several years of deployment, and Ookla’s own framing, treating this as a paradox rather than a verdict, suggests the firm expects VoNR to close the gap as 5G Standalone networks mature further into 2027 and beyond.
Global Call Setup Times: VoLTE vs VoNR by Operator
| Market / Operator | VoLTE Median Setup | VoNR Median Setup | Faster Technology |
|---|---|---|---|
| Global median (Ookla/Omdia) | 1.87 seconds | 1.96 seconds | VoLTE |
| Seoul, South Korea (aggregate) | 1.06 seconds | Slower than all measured VoNR deployments cited | VoLTE |
| KT (South Korea) | 1.21 seconds | 1.47 seconds | VoLTE |
| Singtel (Singapore) | 1.22 seconds | 1.07 seconds | VoNR |
| StarHub (Singapore) | Faster baseline cited | 1.82 seconds (32% slower than VoLTE) | VoLTE |
| LG U+ (South Korea, RootMetrics 2H 2025) | 0.692–0.714 seconds | Not separately reported in this test cycle | VoLTE (fastest overall) |
The spread in that table is the real story. A 0.75-second difference between Seoul’s VoLTE result and the global VoNR median is enormous in call-setup terms, while Singtel’s result shows the gap can flip entirely in a well-optimized market. Reporting on the Ookla data, picked up by outlets tracking the Southeast Asian market, specifically flagged the wide variance in how different carriers in the same country handle 5G voice, underscoring that network engineering choices, not the underlying radio standard, are driving most of the difference.
VoLTE vs VoNR: Feature and Maturity Comparison
| Attribute | VoLTE (4G) | VoNR (5G Standalone) |
|---|---|---|
| Underlying network | 4G LTE with IMS core | 5G Standalone (SA) core and radio |
| Approximate deployment maturity | Roughly a decade of carrier optimization | Still expanding; uneven optimization by market |
| Fallback risk | Can fall back to CSFB on older networks | Can fall back to VoLTE if SA coverage is incomplete |
| Global coverage profile | Near-universal on modern 4G/5G phones | Available in markets including South Korea, China, the US, and Japan per 2026 industry guides |
| Theoretical latency ceiling | Higher than VoNR’s theoretical minimum | Lower, but only reachable with continuous SA coverage |
| Ookla’s current assessment | “Gold standard for consistency and speed in mature deployments” | Competitive or better only where operators have optimized it (e.g., Singtel, Yes 5G) |
What Comes Next: Five Predictions for Mobile Voice
- Carriers in markets where VoNR already beats VoLTE, such as Singapore’s Singtel network, will likely lean harder into 5G voice marketing over the next year, using Ookla-style benchmarks as proof points.
- Operators posting weaker VoNR results, including KT and StarHub in Ookla’s data, are likely to keep VoLTE as the default voice path for most subscribers well into 2027 while they continue tuning standalone 5G voice in the background.
- Expect more independent benchmarking firms to publish operator-level VoLTE-versus-VoNR breakdowns in 2027, since Ookla’s granular, per-carrier data format is likely to be replicated by rivals once marketing departments start citing favorable numbers.
- As device support for VoNR broadens beyond flagship phones, the performance gap documented in October’s report should narrow, following the same multi-year optimization curve VoLTE went through after its own rocky early rollout.
- Enterprise voice and VoIP providers, following Aircall’s current guidance, will likely continue defaulting business-critical calling to 4G LTE rather than 5G voice until VoNR’s consistency, not just its best-case speed, matches what VoLTE already delivers across a wider range of markets.
The Bigger Picture for 5G Rollouts
Ookla’s report does not argue that 5G is a failure, and it is explicit that VoNR already wins in multiple real deployments. What it does argue, in the plainest terms the company has used on this topic, is that network generation alone does not determine call quality. Maturity, optimization, and coverage continuity matter more, which is why a decade-old technology running on 4G still beats a newer one running on 5G in most of the markets Ookla measured in 2026. That is an uncomfortable finding for an industry that has spent years selling 5G as a strict upgrade over everything that came before it, including something as basic as placing a phone call.
It is also a useful reminder for consumers evaluating network claims. A carrier’s 5G coverage map says little about whether a phone call placed on that network will connect in one second or two. Until VoNR’s consistency catches up with VoLTE’s decade of refinement, the practical advice emerging from Ookla’s October 2026 report, and echoed in Aircall’s current guidance to business customers, is that the “newer” network is not automatically the better one for voice.
Frequently Asked Questions
Is 4G actually better than 5G for phone calls?
According to Ookla’s October 2026 report, VoLTE running on 4G beat VoNR running on 5G Standalone networks on call setup time in most markets studied, including a global median of 1.87 seconds for VoLTE versus 1.96 seconds for VoNR. VoNR did outperform VoLTE in several individual markets, including Singapore’s Singtel network, so the result varies by operator rather than being universal.
What is the difference between VoLTE and VoNR?
VoLTE, or Voice over LTE, carries voice calls as IP data over a 4G network. VoNR, or Voice over New Radio, does the same job but entirely within a 5G Standalone network, without using a 4G core at any point.
Why did South Korea post the fastest call setup times?
Ookla’s report credits Seoul’s VoLTE deployment with a 1.06-second median call setup time, faster than any VoNR deployment studied globally, with over 90% of calls meeting a two-second setup target. South Korea’s LTE networks have had roughly a decade of continuous optimization by SK Telecom, KT, and LG U+, which appears to show up directly in the results.
Does this mean 5G networks are worse than 4G overall?
No. The Ookla data is specific to voice call setup time, not overall network speed, data throughput, or latency for other applications. 5G Standalone networks still offer significant advantages for data-heavy tasks; voice calling is simply one area where mature 4G infrastructure is currently holding its own or outperforming newer 5G voice technology.
Can I control whether my phone uses VoLTE or VoNR?
Most phones select automatically based on network availability and carrier configuration, and there is typically no user-facing toggle to force one over the other. Businesses using mobile VoIP platforms like Aircall are advised by the company to favor 4G LTE-only operation in certain scenarios for more consistent call quality.
Which carriers already have VoNR beating VoLTE?
Ookla’s report specifically highlights Singapore’s Singtel network, where VoNR posted a median setup time of 1.07 seconds against 1.22 seconds for VoLTE, and Malaysia’s Yes 5G network, as examples of operators where standalone 5G voice has already matched or exceeded mature 4G voice performance.
Is VoNR expected to improve over time?
Industry guides on the topic describe VoNR’s rollout as expanding but uneven as of 2026, available commercially in markets including South Korea, China, the United States, and Japan. Given that VoLTE itself took years to mature after its early-2010s launch, operators and analysts generally expect VoNR’s consistency to improve as 5G Standalone coverage and device support broaden.
What does call setup time actually measure?
Call setup time measures how long it takes from initiating a call to the connection being established and ready to carry audio. Ookla treats approximately two seconds as the benchmark for a responsive voice experience, and uses the percentage of calls meeting that target as a consistency metric alongside the median setup time itself.