It depends to some extent on experimental parameters. For example, how long can your oligos be? This old publication has a functional site-directed mutagenesis primer with 9 mismatches, but the oligo is significantly longer than the comparator oligos, and it required significantly higher concentrations of $MgCl_2$ (Figure 1). The extra length will buffer the sequence overall against mismatches, as you can imagine, and the higher $MgCl_2$ will lower binding stringency (caveat emptor).
The Addgene blog has the following recommendations for site-directed mutagenesis primers:
As a rule of thumb, 11 bp of complementary sequence on either side of the desired mutation (usually 1-3 mismatched bases) is sufficient for your primers to successfully anneal to the plasmid of interest during the PCR reaction. Ideally, your primers should be free of palindromic and repetitive sequences, but if present, a minor extension can usually ensure that the 3’-base(s) do not form secondary structures. The introduction (or ablation) of a restriction site through mutagenesis vastly facilitates the subsequent process of screening for succesfully mutated clones. Forward and reverse primers are designed to be complementary, but each primer may extend beyond the complementary region as long as an overlap with a minimum 6 bp is maintained. This overlap ensures that the PCR generates a nicked circle rather than a linear product (see figure).
By these parameters, I think that your suggested primer might be ok in theory, but would be more likely to work with additional homology (maybe 2-3bp) upstream and downstream.
It will likely be a finicky reaction as you are pushing beyond the rule-of-thumb limits for mutation size.