"Codes" is actually correct usage and is quite common in academia and in descriptions of commercial products in fields that utilize numerical methods, such as topology optimization or finite element analysis. It is never used to designate a random program that Joe Coder implemented.
What is really meant by this is "well-known and well-tested libraries that do the job as fast as possible". The complexity of the numerical methods is very high, and since performance and accuracy are critical, the demand on the programming skills of their coders is huge. For this reason, the number of numerical method programming libraries is relatively small, and every self-respecting software product in the field uses one or another.
That being said, I would consider any usage outside this very specific context to be incorrect.
Update: By request, some examples below, from different sources:
- Leading Optimization Codes Integrate into Multiphysics and FEA Environments (press release)
- Application of Flow and Transport Optimization Codes to Groundwater Pump and Treat System Optimization (conference proceeding)
- How to Guarantee Finite Termination of Verifying Global Optimization Codes (journal paper)
- A comparison of finite element codes for the solution of biphasic poroelastic problems (journal technical note)
- The usage of standard ﬁnite element codes for computation of dispersion relations in materials with periodic microstructure (journal paper)
Hopefully that's convincing enough. More examples to be had if one searches for "optimization codes", "finite element codes" or similar.