WandB models updated

[koaning]

I noticed that some of the models were out of date, so I updated the list.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
marimo-docs	Ready	Preview, Comment	Jul 1, 2026 11:07am

[Light2Dark]

tq!

[Copilot]

Pull request overview

Updates the curated LLM model catalog for the wandb provider (Weights & Biases) in packages/llm-info, based on the latest WandB inference model list, so downstream consumers (e.g., marimo’s AI model picker/registry) have an up-to-date set of available models.

Changes:

• Added new WandB model entries (e.g., DeepSeek V4-Flash/V4-Pro, GLM-5.2, Kimi K2.7 Code, Qwen3.6 variants, JetBrains Mellum2).
• Removed/replaced several older WandB entries.
• Applied widespread formatting-only changes to cost: inline maps across multiple providers.

[cubic-dev-ai]

1 issue found across 1 file

Architecture diagram

sequenceDiagram
    participant UI as Marimo UI
    participant Backend as Marimo Backend
    participant Provider as W&B Provider Config
    participant Models as LLM Info Registry

    Note over UI,Models: Static model configuration update flow

    UI->>Backend: Request available LLM models for provider
    Backend->>Models: Load models.yml configuration
    Models-->>Backend: Return model definitions
    Backend->>Provider: NEW: Return updated W&B provider model list
    Provider-->>Backend: Model list with updated names/costs
    Backend-->>UI: Return formatted model selections

    Note over UI,Models: Key configuration entries updated in models.yml
    Note over Providers: wandb provider entry updated with latest model versions, costs, and capabilities

Loading

<sub>Reply with feedback, questions, or to request a fix.

Re-trigger cubic</sub>

[Light2Dark]

Filled in missing wandb model metadata in models.yml:

• Added release_date for 8 models that were missing it
• Added cost from W&B inference pricing for those same models
• Normalized all wandb entries to capabilities: [thinking, tool_calling]
• Fixed GLM-5.1 release date (2026-04-07) and gpt-oss pricing to match W&B's current rates