Two groups of chosen models of elastomers networks have been discussed on the basis of literature data and own research results. The first group of models encloses: statistical theory of elasticity, Mooney-Rivlin model, model with constraints on chain and junctions, tube model by Heinrich, Straube and Helmis as well as Tschoegl model. They use simple experimental measurements of stress - strain dependence. However the concentration of chemical crosslinking joints of network, N-c, and crosslinking efficiency E-c, calculated for NR and IR cured with dicumyl peroxide (DCP) are much higher than the values resulting from the known mechanism of these elastomers; curing with a peroxide (E-ct = 2) (Table 1, Figs. 3 and 4). In the second groups of models, by Charlesby-Pinner (concerning the rubber showing M-w/M-n approximate to 2) and by Langley and Pearson (any value of M-w/M-n), there is necessary (for proper analysis) to do labour-consuming determinations of sol content as well as molecular weight and polydispersity of a rubber before curing, but this way determinated values of N-c and E-c, in case of curing NR and IR with DCP [(E-c = 1.8-2.0), agree with theoretical values (Table 1)]. These methods can be recommended to complex analysis of physical structure and topology of a network. In networks of cured elastomers there are both the chemical and permanent topological joints. Their part in total number of network joint reaches up to 80% when chemical crosslinking degree is not high. Key words: networks of elastomers, chemical and topological joints, chain entanglement, comparison of network models, crosslinking efficiency.