From its introduction in 1952 and continuing until the mid 1980s, the 6146 found its way into virtually every manufacturer's line of transmitters. 

There are actually three iterations of the 6146 which are: 6146, 6146A and 6146B. 

The 6146 and 6146A differ in only the heater construction which provides a ruggedisation for successful operation and longevity  in harsh conditions.  
The real difference comes between the 6146(A) and the 6146B which was introduced in 1964 and the B variant had a  redesigned anode construction that allowed anode dissipation to increase from the 6146/6146A limit of 25W up to 35W - a healthy increase of 40%.   Whilst on the subject of 6146B, it is worth mentioning that many users prefer the legendary and sought after "Japanese 6146B", the Matsushita S2001A
 To further compound the story there are also the commercial versions of these valves which were used in commercial FM equipment and the military further ruggedised variant oh, and the 12V heater variant too………..  
Lets take the military variant, the 6146W these can be a a bit of a problem because manufacturers were liable to label either 6146, 6146A or 6146B as….you’ve got it…..6146W but thankfully as large stocks of 6146B occurred and WITH standardisation on usage of the improved device, most 6146W produced after 1966 are actually rebadged 6146B.  
Commercial variants abound, first the was the 6293 with a massive anode and a 1KW dissipation, so rated as the valve was designed for pulse operation, then there is the commercial equivalent of the 6146B - the 8298A - sometimes you see a 6146B cross banded with the commercial notation but additionally, valves showing the commercial notation only exist.    
For the 12V heater variants we have the 6883 which is the 6146 with the higher heater voltage, similarly the 6883A os the 6146A with a different heater and obviously, the 6883B is the 6146B with the higher heater voltage. 
 There is much misinformation on the net about this family of valves but basically there are two things to be mindful of.  Firstly, precise matching for output stages is not really necessary - with +/- 20% giving adequate performance  especially when you consider that most new matched pairs of valves can show variance of this magnitude within a few hours of use.    The second is that whilst it may be OK to use 6146B in place of 6164 and 6146A in a rig designed for the latter it is most certainly not AOK to use 6146 or 6146A in place of 6146B in a rig designed for the latter!