Currently, the stateful tests under tests/stateful/ are not using the latest features of the benchmarking framework:

• They still hardcode gas cost values instead of using bytecode.gas_cost(fork).
• They use the blockchain test wrapper rather than the BenchmarkTest wrapper.
• For CREATE2 memory setup, they don’t use the create2_preimage layout configuration.

Adopting these features would significantly simplify the benchmark tests and improve maintainability.