Why do the other model providers not have a robust json mode yet?

This is a pretty killer requirement for doing anything serious with LLM's i.e. recieving structured output enables all kinds of workflows beyond just chatbots.

Can we get reliable structured output from Anthropic or anywhere else? I saw this https://docs.anthropic.com/en/docs/test-and-evaluate/strengthen-guardrails/increase-consistency but it looks like just prompt engineering rather than a specific API level change which is what I assume OpenAI does, something like switching over to a fine-tuned version of the model designed to output only JSON.