[Roadmap] Update CTBenchmarks for OptimalControl v2 and benchmarking paper

## Objective

Update CTBenchmarks.jl to be compatible with OptimalControl v2 and prepare a comprehensive benchmarking paper.

## Context

CTBenchmarks.jl needs to be updated to work with the new OptimalControl v2 API. This includes updating problem formulations, solver interfaces, and benchmark infrastructure.

## Why

Benchmarking is essential to validate the ecosystem and compare performance against other tools. A comprehensive benchmarking paper will showcase the capabilities of the control-toolbox ecosystem and provide reference performance data.

## Scope

- Update for OptimalControl v2 compatibility
- Benchmark NLP solvers, modelers, linear solvers, AD backends
- Compare against CasADi, InfinitOpt, JuMP, Grape
- Write a comprehensive benchmarking article

## References

- [CTBenchmarks.jl](https://github.com/control-toolbox/CTBenchmarks.jl)
- [Roadmap comment](https://github.com/orgs/control-toolbox/discussions/68#discussioncomment-17584843)