Reporting snapshot · 30 September 2026. The figures come from Opensignal's Pakistan Mobile Network Experience report for the second quarter of 2026, published on 30 September and based on crowdsourced quality-of-experience measurements collected over a 90-day window in Q2 2026. The Express Tribune's business desk carried the report's headline numbers and operator-by-operator breakdown; the figures below are drawn from that wire-style account and from Opensignal's methodology note on crowdsourced scoring. The report itself is not a regulator-issued document; the Pakistan Telecommunication Authority (PTA) does not publish a parallel 4G benchmark. The crowdsourced sample covers national averages and 20 cities across indoor and outdoor, peak and off-peak measurements.
What happened
Opensignal’s Pakistan Mobile Network Experience report for Q2 2026, published on 30 September, records a small improvement in raw download speed and a much larger deterioration in the everyday video experience on Pakistan’s mobile networks. National 4G download speed rose from 18.6 Mbps in Q1 to 18.9 Mbps, the Express Tribune’s business desk reported on 30 September, citing Opensignal directly. High-end phones averaged 31.3 Mbps, up from 30.2 Mbps.
The gains end at the speedometer. The Opensignal video score — which measures how a stream actually looks and whether it stalls — dropped from 44.97 to 37.53 on a 100-point scale. Upload fell from 5.81 Mbps to 5.32 Mbps. Time-on-network moved marginally, from 94.21 per cent to 93.67 per cent of connected time on 4G.
The same combination — flat speed, falling video — repeats at every operator. No operator escaped the video drop. Jazz’s national video score went from 47.00 to 41.51, Zong’s from 47.06 to 41.66, Ufone from 42.33 to 36.64. Telenor — listed separately although it was merged into Ufone on 1 July 2026 — dropped from 34.84 to 30.30. The crowdsourced dataset covers national averages and 20 cities, and is not a regulator-issued compliance table.
Speed barely moved, video fell on every operator
The two operators that score highest on raw speed did so by small margins. Jazz was at 17.62 Mbps in Q1 and 17.61 Mbps in Q2; Zong was at 18.62 Mbps in Q1 and 18.55 Mbps in Q2. Telenor improved from 6.66 Mbps to 7.50 Mbps and Ufone from 12.36 Mbps to 12.82 Mbps, both from a lower base. A 0.07 Mbps difference at the top is well inside the report’s measurement noise.
The seven-point video-score fall with flat speeds is the headline finding: a file can finish downloading while a live stream still buffers, the Tribune paraphrased Opensignal’s interpretation, pointing to congestion, uneven capacity at busy hours or how traffic is handled rather than a gap on the coverage map. The report also separates device effects from network effects — high-end phones averaged 31.3 Mbps versus 18.9 Mbps on all devices — a 12 Mbps gap built into the same physical networks.
Zong led download, Jazz led video in the capital
City rankings do not follow the national table. Jazz took download in Islamabad at 20.86 Mbps, with the strongest large-city video score at 50.22, and led Rawalpindi, Peshawar, Faisalabad, Sukkur, Hyderabad and Swabi. Quetta was a standout: Jazz’s 22.41 Mbps there is one of the highest city prints in the report.
Zong led Lahore at 21.26 Mbps (video 48.89, games 53.56, time-on-network 98.18 per cent), plus Multan, Karachi at 18.27 Mbps, Gujranwala, Sialkot, Chakwal, Khairpur and Gwadar. In the three northern cities where the Special Communications Organisation is also measured — Mirpur, Skardu and Muzaffarabad — Zong led download in each.
Karachi splits the spoils: Zong fastest, Jazz best video at 47.18, Ufone best games at 56.03. Telenor’s 12.19 Mbps was its least-bad large-city download and still last. Ufone’s games lead holds in most printed cities; its coverage is thinner in places, with time-on-network at 87.97 per cent in Hyderabad, 85.4 per cent in Khairpur and 48.67 per cent in Skardu. Telenor’s weakest city print is Gwadar: 3.87 Mbps download, 1.27 Mbps upload, video 13.75 and time-on-network 59.18 per cent.
How the four operators compare on national metrics
On national 4G, Zong led download at 18.55 Mbps, upload at 7.81 Mbps and time-on-network at 96.59 per cent, the report’s highest availability reading. Its video score of 41.66 was a shade ahead of Jazz; games were 47.28, second to Ufone. Jazz was second on download at 17.61 Mbps and upload at 6.26 Mbps, with games 40.15, its weakest metric and the lowest of the four.
Ufone did not win on speed. It won on games at 56.03 nationally, with download 12.82 Mbps, upload 4.42 Mbps, video 36.64 and time-on-network 92.16 per cent — the lowest availability. Telenor finished last on every metric except availability: download 7.50 Mbps (less than half of Zong’s), upload 2.82 Mbps, video 30.30, games 37.88. Time-on-network at 92.58 per cent was close to the pack. The problem was quality when connected, not whether a signal was there.
How this fits Pakistan’s broader digital push
The Opensignal figures land in a year that has pushed Pakistan’s digital economy onto a higher political priority. The IT ministry’s 28 September request to Huawei to set up an R&D centre in Islamabad was framed around ICT-enabled services exports, and the 22 September USEFP Innovation Hive featured Pakistan’s emerging ICT export pitch to a Chargé d’Affaires audience. Cybercrime enforcement — the 22 September NCCIA raids that detained 212 people in three illegal Islamabad call centres — has been the most-covered Tech-and-telecom story of the last 60 days.
Inside that frame, the Opensignal report is a reminder that the export story sits on top of an uneven domestic experience. Jazz and Zong can both put 17–19 Mbps download averages in front of investors; the same two operators posted national video scores below 42, meaning live-streamed classrooms, news clips and shop fronts arrive patchily on the phones in readers’ pockets. The PTA has not published a parallel benchmark; the regulator’s role is licence and spectrum compliance rather than consumer experience scoring, which is the gap Opensignal’s crowdsourced methodology fills.
What is still uncertain
Three limits follow from the available reporting. First, Opensignal’s Q2 numbers are crowdsourced, not a regulator-issued compliance table; they describe what users saw between April and June, not what networks were licensed to deliver. Second, Telenor is reported as a separate operator in the Q2 dataset even though the Telenor–Ufone merger completed on 1 July 2026; Q2 readings reflect the standalone Telenor network, and post-merger figures will not appear in this format. Third, the device-effect gap is built into the same physical networks and tells users that handset choice, not just SIM choice, decides how fast a Pakistan 4G connection feels.
The seven-point video-score fall with flat speeds is the open question. If a future quarter restores video while speeds hold, the report will read as a network-policy fix. If speeds climb and video keeps falling, the gap will read as a capacity constraint rather than a one-quarter congestion episode.
Sources & reporting notes
This is a synthesis of published reporting on the Opensignal Q2 2026 Pakistan Mobile Network Experience report and its operator breakdown. Sources were reviewed on 30 September 2026.
- The Express Tribune — Faster phones, worse video on 4G30 September 2026 · Usman Hanif’s wire-style summary of the Opensignal Q2 2026 Pakistan Mobile Network Experience report; carries the national averages, the operator-by-operator table, the city-level breakdown and Opensignal’s congestion-versus-coverage interpretation.
- Opensignal — Market Insights30 September 2026 · The primary record for the Q2 2026 Pakistan Mobile Network Experience report; carries the methodology note on 90-day crowdsourced scoring, indoor and outdoor coverage and the 20-city national sample. Used here to confirm that the report is crowdsourced rather than regulator-issued.
- The Express Tribune — Telenor Pakistan, Ufone formally merge after IHC approval1 July 2026 · Independent Tribune coverage of the 1 July merger of Telenor Pakistan into Pak Telecom Mobile Limited; used to confirm why Q2 2026 readings still treat Telenor as a separate operator.
- The Express Tribune — Huawei asked to set up R&D centre29 September 2026 · Independent context on the IT ministry’s 28 September pitch to Huawei, used to place the Opensignal figures inside the broader digital-economy push.
