Add Runcrate as an LLM provider #13419

Runcrate (https://www.runcrate.ai) serves open models (DeepSeek V4.1 Flash, GLM-5.3, Qwen3.8 Flash, gpt-oss 120B and more) through an OpenAI-compatible API at https://api.runcrate.ai/v1. Today users reach it with provider: openai and a custom apiBase. I'd like to add a runcrate provider the way CONTRIBUTING describes (a BaseLLM subclass on the OpenAI class, the LLMs array, a docs page under customize/model-providers/more, and an entry in the Add Model UI), the same shape as the Mimo provider. I have it working locally and can open the PR.

Disclosure: I work on Runcrate.