Do you guys model 208V lighting panel breakers? I'm a bit confused on whether it is necessary to model the main breaker inside a lighting panel.

Maybe. Depends on the transformer size feeding it (<125 kVA, single transformer, <240 VAC according to current IEEE 1584 standard). NFPA 70E-2009 quoted this cutoff following IEEE 1584. In 70E-2012, it is dropped in anticipation of IEEE 1584 being updated to say something different.

Some research has been done which shows that arcs are sustainable at this voltage. The next edition of IEEE 1584 is likely to contain verbiage which changes the cutoff values. The cutoff is likely to be voltage and current dependent.

Note that panels from Square D (and maybe others) are often labelled as "no arc hazard" because the manufacturer has tested the panel and determined that their dead front panels have no arc flash hazard. Not just "PPE 0", but nada, none, zero, zilch. As in go ahead and wear your polyester suit while resetting breakers with the cover on.

Wow, nothing like spreading false information.

I am not aware of a single manufacturer of lighting panels, particularly Square D, that says 'no arc flash hazard' based on their testing.

Containing an arcing fault internal to the lighting panel is not the same issue as not causing an arc fault.

I do know manufacturers that make statements based on the testing done by UL in order to get an AIC rating for their protective devices and an SCCR for their enclosures, (i.e. that cotton surrounding a Type 12 enclosure did not ignite when the devices interrupted current at their full AIC). They often interpret these results as saying normal operation, such as On-Off switching but not resetting, of a device does not present an increased hazard of causing an arc flash. They all go on to say that PPE selection must be done per the employer's Electrically Safe Work Practices program

By the same token, I am not aware of any manufacturer of lighting panels that says an internal main breaker will mitigate the incident energy inside of the panel during an arcing fault.

So to the OP's original question - there is usually no need to model the contribution of a panel's main breaker. You simply model up to and including its line side lugs.

We worry less about the main breaker that the incomming leads. Usually the incomming side has a far higher category than after the main. The only reason we look at the main is to calculate something down stream. The probability of something going wrong after the main is greater than ahead of the main but there has been some research done on propagation - same with under 240V and 125kVA. I think the jury is still out but I lean on the conservative side.

In our studies, and those performed for our company, we require all panelboards to be modeled, except for small single phase PBs. the software we use (EasyPower) is a living document of our facilities and as our intent is to keep the studies up to date by updating all of our panelboards, switchboards, etc. EP does have a limit, but a "workaround", for single phase, but we're not expending too much effort there.