Old query, jsonb:
                                                            QUERY PLAN
----------------------------------------------------------------------------------------------------------------------------------
 Nested Loop  (cost=0.48..1359.15 rows=2714 width=168) (actual time=0.067..41.844 rows=2476 loops=1)
   ->  Seq Scan on users  (cost=0.00..40.14 rows=2714 width=309) (actual time=0.019..1.851 rows=2476 loops=1)
   ->  Aggregate  (cost=0.48..0.48 rows=1 width=32) (actual time=0.015..0.016 rows=1 loops=2476)
         ->  Function Scan on jsonb_each_text tmp  (cost=0.00..0.48 rows=1 width=32) (actual time=0.013..0.014 rows=1 loops=2476)
               Filter: ((val = 'true'::text) AND (col = ANY ('{premium,pepsi,spawnmason,staff,developer}'::text[])))
               Rows Removed by Filter: 13
 Planning Time: 0.156 ms
 Execution Time: 42.072 ms
(8 rows)

Your first suggestion, with VALUES:
                                                    QUERY PLAN
-------------------------------------------------------------------------------------------------------------------
 Seq Scan on users  (cost=0.00..87.64 rows=2714 width=168) (actual time=0.026..7.057 rows=2476 loops=1)
   SubPlan 1
     ->  Values Scan on "*VALUES*"  (cost=0.00..0.02 rows=2 width=32) (actual time=0.000..0.002 rows=1 loops=2476)
           Filter: column2
           Rows Removed by Filter: 4
 Planning Time: 0.146 ms
 Execution Time: 7.166 ms
(7 rows)

Your last suggestion, with string_to_array:
                                               QUERY PLAN
--------------------------------------------------------------------------------------------------------
 Seq Scan on users  (cost=0.00..42.86 rows=2714 width=168) (actual time=0.021..1.816 rows=2476 loops=1)
 Planning Time: 0.071 ms
 Execution Time: 1.945 ms
(3 rows)