RAN1, the 3GPP working group that designs the physical layer of the radio interface, meets in Calgary from 16 to 20 November 2026, in the same week as SA2 and 13 other working groups. Most of its agenda belongs to the Rel-20 study of the 6G radio interface (6GR). A study is the stage in which options are assessed and narrowed before binding specifications are written; its results are collected in TR 38.760-1, which enters the meeting as v0.8.0 (R1-2610402). In Jeju in October RAN1 settled many details but moved the larger radio choices to this meeting, as our summary of RAN1#126-bis described. This preview is based on the documents submitted for Calgary, read before the meeting starts. Nothing in it has been decided yet.

Calgary is the last meeting for decisions

The rapporteurs' workplan (R1-2610401) asks RAN1 in Calgary to "draw conclusions" on every part of the 6G radio. For its next meeting, RAN1#128 in Osaka on 25–29 January 2027, the same plan says only "Summarize the outcome". The TR editor writes that "the RAN1 study is expected to be completed at the January meeting". The vice-chair's guidance for this week asks delegates to look for a direction in the first session on each topic and, in the second and third, to try to make "the decision" according to the work plan. Of the 1,869 documents in the meeting's TDoc list, 1,322 belong to the 6G study.

Decisions expected this week. These are the choices RAN1 explicitly deferred from Jeju to Calgary, plus the conclusions the workplan names:

  • the default period of the synchronization signal block (SSB) for initial cell search: 20, 40, 80 or 160 ms;
  • the length of the SSB: 4, 5 or 6 OFDM symbols;
  • whether one transport block may span a 400 MHz carrier, or two blocks of up to 200 MHz are used;
  • how codewords map to MIMO layers in the downlink;
  • the maximum polar mother code length for large downlink control messages (512 or 1024), and whether the 5G Reed-Muller code for small uplink control messages has a drawback that justifies replacing it;
  • the cell-wide number of wake-up signal sequences (32, 64, 128 or 256);
  • "definite conclusions" on the benefit of spectral extension or truncation for the uplink waveform, and the remaining details of the new LDPC base graph and of uplink 1024QAM.

What moves elsewhere or later.

  • To RAN4. Whatever SSB period RAN1 picks, the performance requirements behind it are for RAN4 to define. The Jeju agreement says so explicitly, including the assumptions about how phones combine repetitions. RAN4 also specifies how uplink transmit filters are characterized, for example through a spectrum mask.
  • To RAN2. System information, paging, random-access procedures and the phone capability framework are RAN2 topics. RAN2's part of the study runs until May 2027.
  • To the TR, after January. The TR editor plans an email discussion after the January meeting on how to capture the study's final outcome, "not to reopen the previous agreements".
  • Beyond Rel-20. The TR states that a 400 MHz uplink "may be revisited in future releases beyond Release-20". The Jeju agreement on the SSB period adds that the detailed phone capability "will be discussed separately".

Our reading. If RAN1 does not converge on a question this week, the workplan leaves no meeting dedicated to it. It would have to be squeezed into the January summary or carried into the normative phase as an open point.

How long may a 6G base station stay silent?

A phone that switches on looks for the SSB, a short burst with the primary and secondary synchronization signals and the broadcast channel. In 5G NR a phone searching for a cell assumes the SSB repeats every 20 ms. The longer the gap, the longer the base station can sleep, which is the core of 6G network energy saving. The cost falls on the phone: it waits longer for a burst, or it has to combine repetitions to reach the same coverage.

The Jeju agreement lists five options: 160 ms with up to four repetitions inside one period (Option 1), two 80 ms variants (2a, 2b), two 40 ms variants (3, 4), and 20 ms with combining over four periods (Option 5). The Calgary contributions fall into three groups:

  • 160 ms, or at least 80 ms. Ericsson (R1-2609297), Huawei (R1-2609309), Samsung (R1-2609601), ZTE (R1-2609246), KDDI (R1-2610617), Futurewei and ETRI support 160 ms. CATT prefers 80 ms and accepts 160 ms (R1-2609671). AT&T asks RAN1 to deprioritize the 40 ms options (R1-2610320). Its argument: under the base-station power models, the shortest transition from deep sleep to active is 50 ms, so moving from 20 ms to 40 ms is not well motivated.
  • 40 ms. MediaTek (R1-2610228), Xiaomi (R1-2609363), OPPO (R1-2609915), Honor, Jio and Spreadtrum support a 40 ms period, in most cases with two repetitions (Option 4). Apple prefers 40 ms to 80 ms with four repetitions, which it says doubles both the overhead per detection attempt and the minimum acquisition time (R1-2609855). Nokia suggests choosing between 40 ms and 80 ms with two repetitions (R1-2609078). NTT DOCOMO wants to keep studying 160, 80 and 40 ms; its own simulation at 7 GHz shows the 40 ms option performing better than 160 ms (R1-2610413).
  • Keep 20 ms. Qualcomm supports the 5G value of 20 ms and asks that connected phones get a 20 ms SSB whatever default is chosen (R1-2610468). Google prefers 20 ms, or at most 40 ms (R1-2610084). Nordic Semiconductor and Itron want 20 ms for the coverage layer below 3 GHz (R1-2610079).

Our reading. The line runs roughly between companies that build and operate base stations and companies that build phone chips and phones, though not cleanly: Samsung, which makes both, supports 160 ms, Nokia, a base-station vendor, suggests 40 or 80 ms, and DOCOMO, an operator, has published results in favour of 40 ms. Many contributions on all sides add the same mitigation: a sparser synchronization raster, the set of frequencies a phone has to scan, which cuts search time whatever period is chosen.

How long is the burst?

RAN1 agreed in Jeju that one 6G SSB spans 4, 5 or 6 OFDM symbols in a slot. With four, the primary and secondary synchronization signals get one symbol each, as in 5G; with five or six, the primary signal gets two, which helps detection at the cell edge. Four symbols are supported by ZTE, vivo, Xiaomi, Qualcomm, ETRI, Honor, Jio, Sharp and others. Six are supported by Ericsson (which also wants the primary signal and broadcast channel side by side in frequency), LG Electronics, CATT and Sony. OPPO and IIT Madras propose five. Nokia and Futurewei accept five or six, and Google four or five. The choice interacts with the period: a longer burst gives each detection attempt more energy, which matters more if bursts are rare.

One block of data or two above 200 MHz?

Around 7 GHz, the TR already provides for phones that receive on a carrier of up to 400 MHz in the downlink, as an optional capability, and transmit on up to 200 MHz in the uplink. RAN1 has to decide whether one transport block, the unit of data the physical layer codes and acknowledges, may span the whole 400 MHz (Option A), or whether an allocation between 200 and 400 MHz carries two blocks side by side, each up to 200 MHz (Option B). Jeju moved the choice to Calgary with complexity, power consumption and performance as criteria.

  • One block up to 400 MHz: Ericsson (R1-2609544), NTT DOCOMO (R1-2610417), China Mobile (R1-2609415), ZTE, NVIDIA (R1-2610254), ETRI, Sharp and Jio. Qualcomm's comparison concludes that Option A "greatly simplifies UE processing" and that Option B shows no performance advantage (R1-2610472).
  • Two blocks of up to 200 MHz: Huawei (R1-2609313), MediaTek (R1-2609206), Apple (R1-2609859), Intel, Google, Xiaomi, LG Electronics, Honor and InterDigital. Several of them want a separate modulation and coding scheme for each block.

Unlike the SSB period, this question does not split along the usual lines: Qualcomm and MediaTek, both phone-chip makers, are on opposite sides, and so are Ericsson and Huawei.

Smaller decisions

  • Codes for control messages. For downlink control messages larger than the 5G maximum of 140 bits, several contributions, including vivo, NTT DOCOMO and Qualcomm, keep the 5G mother code length of 512; China Mobile raises 1024, and CATT supports it for better control-channel coverage. For uplink control messages of 3 to 11 bits, Samsung says no drawback of the 5G Reed-Muller code has been shown. Google and Jio point to an error floor for some combinations of payload and length, and with Qualcomm and Ericsson propose keeping the Reed-Muller code with improved rate matching.
  • Wake-up signal. The most common explicit proposal for the number of sequences is 64; Samsung and InterDigital want 256, ZTE 128, Qualcomm and Honor 32 or 64.

Next door in RAN2: early channel feedback for 5G-Advanced

One 5G-Advanced question from our list of open decisions sits with RAN2, which also meets in Calgary. Under the Rel-20 NR MIMO Phase 6 work item, the fourth message of random access can already ask a phone moving from idle to connected mode to report channel state information or send sounding signals, so the network can schedule it well from the start. The configuration has to reach the phone in system information before it sends its third message. RAN1 left the choice of system information block to RAN2 (R2-2608113), and RAN2 already settled it in October: "One new SIB is introduced" (quoted in the rapporteur's open-issue list, R2-2609142). What is left for Calgary is whether SIB1, which every phone reads, also says what the new block contains, so that a phone without the feature can skip it. MediaTek (R2-2609143), Nokia and ZTE propose such an indication; Qualcomm (R2-2608917), Ericsson, Huawei, LG Electronics, OPPO, CATT, Ofinno and Lenovo see no need for it.

The 6G radio as a whole

Put together, TR 38.760-1 v0.8.0 and this week's agenda give a fairly complete picture. The middle column is what the TR already records; the right column is what remains open, with the items due in Calgary marked.

AreaAgreed so farStill open
WaveformCP-OFDM in the downlink; CP-OFDM and DFT-s-OFDM in the uplink, both as in 5G NR. Spectrum shaping for π/2-BPSK; no low-PAPR scheme for 16QAM.Spectral extension or truncation (Calgary).
Frame and numerologyNR frame as the starting point: 10 ms frame, 14-symbol slot, normal cyclic prefix. 15 kHz (FDD) or 30 kHz (TDD) below 6 GHz, 30 kHz around 7 GHz, 120 kHz at 24–52 GHz.How dynamic TDD decides the link direction; several options remain in the TR.
BandwidthAround 7 GHz up to 400 MHz downlink (optional) and 200 MHz uplink. The simplest devices support 20 MHz of RF bandwidth.One or two transport blocks above 200 MHz (Calgary).
Coding and modulationLDPC for data with a new base graph (BG3) for rates beyond 5G; polar codes for control above 11 bits. Up to 1024QAM in the downlink, 1024QAM in the uplink for fixed wireless access.Polar mother code length, Reed-Muller for 3–11 bits, BG3 details (Calgary).
Cell search and accessTwo SSB positions per slot; up to 4 or 8 beams below 3 GHz, 8 below 6 GHz and 16 around 7 GHz. Four-step random access as the baseline.SSB period and SSB length (Calgary).
DuplexingSub-band full duplex at the base station, configured semi-statically.–
Energy savingA one-symbol wake-up signal of at most 10 MHz, monitored with the paging period in idle mode.Number of wake-up sequences per cell (Calgary).
SatellitesPhones with GNSS, degraded GNSS or no GNSS in the same cell; pre-compensation from TR 38.742 as the starting point.Conclusions on the NTN candidate technologies (Calgary).
AI/ML and sensingLists of use cases and evaluation methods; CP-OFDM as the starting point for sensing.The workplan asks only to "converge to draw conclusions" on sensing.

Our reading. The 6G radio is taking shape as a direct descendant of 5G NR: the same waveforms, the same frame, the same families of codes. The differences sit where the study's targets point. Energy saving is the main one, with longer SSB periods and a wake-up signal. Wider carriers at 7 GHz and a code for higher data rates come next, followed by satellites designed in from the start. The SSB period decided this week is the single choice that most changes how a 6G network behaves compared with 5G. AI/ML and sensing, often presented as the defining features of 6G, are the least settled parts of the radio study.

What to watch

  • Whether RAN1 picks one SSB period or only narrows the list. A narrowed list would push the choice into the January summary meeting.
  • Whether a long period comes with a sparser synchronization raster, which needs input from RAN4.
  • Whether the 400 MHz question is settled or split, for example with Option A as the baseline and two blocks as a phone capability.
  • Timing. Under the draft schedule the SSB is discussed online from Monday morning and returns several times during the week. The end-of-meeting chair notes are the first record of what was agreed.

SA2 meets in the same week with its own deadline on the overall 6G architecture (see our SA2#178 preview). The outcomes go to the plenaries in Boston: CT#114 on 7–8 December, RAN#114 on 7–10 December and SA#114 on 8–11 December.

Sources

RAN1#127 agenda, pre-meeting chair notes and session schedule (v02, 11–12 November 2026), the vice-chair's session guidance, the TDoc list of 13 November, the workplan, TR 38.760-1 v0.8.0 with the editor's cover note, and the contributions linked above, read on 15 November 2026. The feature leads' summaries, which usually tally company positions, are not yet available; the groupings above are our reading of the individual contributions and name companies only where their proposal is explicit. RAN2#136 agenda and contributions under agenda item 8.10 (NR MIMO Phase 6). Meeting dates are from the 3GPP meeting calendar. Positions are what companies propose, not 3GPP decisions.