There's many different voltage, and it varies depending of frequencies
From the IEEE 802.3-2008 document borrowed from the official IEE 802.3 get page - but it seems it is no more freely available.
18.104.22.168 AC common-mode output voltage
The magnitude of the ac component of the common-mode output voltage of the driver, measured between the midpoint of a test load consisting of a pair of matched 39 > Ω ± 1% resistors and circuit VC, as shown in Figure 7–13, shall not exceed 2.5 V peak from 30 Hz to 40 kHz and 160 mV peak from 40 kHz to BR.
22.214.171.124 Differential output voltage, open circuit
The differential output voltage into an open circuit, measured at the interface connector of the driving unit, shall not exceed 13 V peak.
126.96.36.199 DC common-mode output voltage
The magnitude of the dc component of the common-mode output voltage of the driver, measured
between the midpoint of a test load consisting of a pair of matched 39
Ω ± 1% resistors and circuit VC, as shown in Figure 7–13, shall not
exceed 5.5 V.
Newer standards may have different limits, but since they are usually retro-compatible, they must not have changed much. I dug trough a few of them but didn't found specifications that contradict those above.
The 802.3bq-2016: Physical Layer and Management Parameters for 25 Gb/s and 40 Gb/s Operation, Types 25GBASE-T and 40GBASE-T document has the table:
in which you can find a Differential-mode voltage of
< 2.4 + 19.68 (f / 30) mVpp close enough to 2.5V.
(Vpp stands for the voltage differential peak-to-peak, i.e. difference between the lowest and highest voltage in a period.)
There's another voltage involved, when using
Power Over Ethernet (POE), which range from 37 to 57V.