How should I evaluate AI coding agents and models for MathWorks Agentic Toolkits?

I want to use MATLAB Agentic Toolkit, Simulink Agentic Toolkit, or Polyspace Agentic Toolkit in my engineering workflows.

Does MathWorks recommend a specific AI coding agent, underlying generative AI model, or AI service for these workflows?

What criteria should I use to evaluate the available options?

Answer 1

[MathWorks Support Team]

Selecting a solution involves both the AI coding agent and the underlying generative AI model or service.

Refer to the following pages for current toolkit requirements and supported platforms:

• MATLAB Agentic Toolkit requirements
• Simulink Agentic Toolkit supported platforms
• Polyspace Agentic Toolkit supported platforms

MathWorks does not recommend or rank specific AI coding agents, generative AI models, or AI services for individual workflows.

The most suitable AI coding agent and model depend on the intended workflow and might differ across workflows. Evaluate candidates using small, representative tasks for each MATLAB, Simulink, or Polyspace workflow you need to support. Use comparable inputs and conditions for each candidate.

Consider the task results together with publicly available information such as coding and reasoning benchmarks, context-window size, cost, and applicable security requirements.

General benchmarks and specifications do not necessarily reflect performance for MATLAB-, Simulink-, or Polyspace-specific workflows.