6AQ5 is a small/cheap 6V6. 6.3V 0.45A heater, 12W plate dissipation. Surely not a Philips invention.
EL95 is new to me. At a glance, 0.2A heater and 6W plate means it is only half the brute that a 6AQ5 is. US-registered as 6DL5.
It's like putting a 215 CID (3.5L) engine in for a 427 CID (7L) engine. It won't go so good.
And Mu of 6V6/6AQ5 is ~~10, for EL95 it says 17. It will need very different bias.
So like the throttle lever comes out with different leverage; you will have to re-adjust the idle and WOT settings.
I am much too lazy to compare pin-outs, but I suggest you triple-check that.
> swappable in a circuit?
EL95 in a hole made for self-bias 6AQ5 would, by coincidence, bias-up to a tolerable DC current. The higher Mu would force less current for the same voltage and bias resistor, roughly half which suits that half-size cathode and plate. If you keep the load impedance the same, the power output is a quarter of the 6AQ5's mighty 4.5 Watts (half current times half voltage is quarter power). If you double the load impedance, you get about half of a 6AQ5's output. The sheet claims 3 Watts out (2/3rd of the 6AQ5 claim) but at an unusually high 12% THD-- heavily fudged.
EL95 is interesting where a 6V6/6AQ5 is overkill, assuming you can give it a hi-Z load or tolerate less than max-possible power. 6V6 has the advantage that you can get a replacement on Friday night, but I guess 6AQ5 is no easier to source than EL95. They run about the same price (~~$10) at the major bottle-stores.