Aha, I take it you are running open carbs right ? Or throttle bodies if EFI now, not sure.
Either way, then without a steady air supply when testing on a dyno everything can be fine, but once moving, the placement of components & the fairing can affect the air pressure around the intakes, as can a cross wind. It's one of the reasons most ramair systems are now centrally mounted on the fairing to reduce the crosswind effect on a dual intake system.
I'd advise you look at a means to guarantee static ambient or higher pressure air at the intake to stabilise the feed their as much as possible.
Re the O2 sensors, they are not so expensive now, the newer Bosch LSU 4.9 is an amazing piece of kit. I have one that I used to manually tune my VFR800 to a new to me TBR performance exhaust on the road & was impressed when on the dyno the AFR was near perfect
If you are worried about tip melt if fitted 12" from the exhaust port, then do what they do in dyno's add a small 8mm port & a vacuum pump, you could use an intake tap vacuum pump to reduce complexity & mount the sensor remotely. If you did two of those, then you could wire in an ignition kill for a set AFR like 15-1 etc, thus saving your engine !
Have a look at outboard engine fuel pumps for an idea as to how an intake vacuum pump works or speak nicely to your dyno guy to let you see how DJ do it for their AFR probe !
Last idea, use surface discharge plugs they are now available quite cheaply, they have no protruding electrode to over heat !
Just a few ideas, hope they help.