DUDES / MODELS

The best tool. For this job.

Model loyalty makes no sense. We select models for the task, the data constraints and the quality of the result.

01

OpenAI · Anthropic · Google

GPT, Claude and Gemini model families are part of the toolbox. This is not a claim that one model wins every benchmark or a fixed list of versions.

02

How we choose

We consider coding quality, reasoning, tool use, context, latency and cost. Security requirements and data retention can rule out a model before performance enters the discussion.

03

More than a chat window

Coding agents, repository context, browser checks, static analysis and CI/CD connect model output to a real engineering workflow.

04

Independent review

A second model can help uncover a missing assumption. Agreement between models is not proof: executable checks and human review still decide whether the result is acceptable.

Got a “could you build…”?

Tell us what to build