This paper introduces a model of quantum computation using local fermionic modes (LFMs), which are sites that can be either empty or occupied by a fermion. The authors show that simulating one fermionic gate can be done using O(m) qubit gates, but with different encodings, this can be reduced to O(log m) or even a constant number of qubit gates. They also find a universal set of fermionic gates and study computation with Majorana fermions, which are halves of LFMs. The paper discusses the relationship between fermionic and qubit quantum computation, showing that fermionic gates can be simulated using qubit gates and vice versa. The authors also show that nearest-neighbor fermionic gates on a graph of bounded degree can be simulated at a constant cost. They further demonstrate that fermionic computation models do not generate new computational classes but provide new descriptions of the standard class BQP. The paper concludes that alternative quantum computation models may be useful for finding new quantum algorithms, error-correcting codes, and fault-tolerant procedures.This paper introduces a model of quantum computation using local fermionic modes (LFMs), which are sites that can be either empty or occupied by a fermion. The authors show that simulating one fermionic gate can be done using O(m) qubit gates, but with different encodings, this can be reduced to O(log m) or even a constant number of qubit gates. They also find a universal set of fermionic gates and study computation with Majorana fermions, which are halves of LFMs. The paper discusses the relationship between fermionic and qubit quantum computation, showing that fermionic gates can be simulated using qubit gates and vice versa. The authors also show that nearest-neighbor fermionic gates on a graph of bounded degree can be simulated at a constant cost. They further demonstrate that fermionic computation models do not generate new computational classes but provide new descriptions of the standard class BQP. The paper concludes that alternative quantum computation models may be useful for finding new quantum algorithms, error-correcting codes, and fault-tolerant procedures.