[Runtimes] Add GLM, Devstral, and MiniMax manifests

[YouNeedCryDear]

PR Description

What this PR does

• adds ClusterBaseModel manifests for GLM, Devstral, and MiniMax
  model variants
• adds ClusterServingRuntime manifests to support those new models
• adds sample InferenceService manifests for the new model entries
• updates the runtime kustomization to include the new runtime
  resources

Why we need it

This expands OME's built-in model catalog and runtime coverage for
GLM, Devstral, and MiniMax variants, and provides sample manifests to
make deployment and validation easier.

Fixes #

How to test

• review the new model manifests under config/models
• review the new runtime manifests under config/runtimes
• review the new sample ISVC manifests under config/samples/isvc

Checklist

• Tests added/updated (if applicable)
• Docs updated (if applicable)
• make test passes locally

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[chatgpt-codex-connector]

💡 Codex Review

https://github.com/sgl-project/ome/blob/dcacbfbe20832beee8c97d959b02925b4e8fc06c/config/models/MiniMaxAI/MiniMax-M2.yaml#L1
Keep MiniMax-M2 path aligned with model kustomization

This rename moves MiniMax-M2.yaml out of config/models/minimax, but config/models/kustomization.yaml still references minimax/MiniMax-M2.yaml, so any workflow that builds/applies the model kustomization will fail with a missing-file error and block publishing the model catalog. Please either keep the old path or update the kustomization entry to the new MiniMaxAI/MiniMax-M2.yaml location in the same change.

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".