chore: replace Qwen3 Coder with Kimi K2.5

[AnthonyRonning]

• Replace Math/Coding default model with kimi-k2-5 (256k, vision-capable)
• Add alias/migration so stored qwen3-coder-480b resolves to kimi-k2-5
• Update marketing badges (remove Qwen3 Coder; keep Qwen3-VL)

---

Summary by CodeRabbit

• New Features

  • Introduced Kimi K2.5 AI model featuring advanced vision recognition capabilities, extended 256K token limit, and Pro-tier subscription access.

• Updates

  • Refined AI model labeling and categorization system for improved user experience.
  • Transitioned math category to utilize the new Kimi K2.5 model as default.
  • Removed previous specialized math model from user-accessible options.

[coderabbitai]

📝 Walkthrough

Walkthrough

Replaces qwen3-coder-480b with kimi-k2-5 across frontend configuration and UI labels, adds aliasing so the old model ID maps to the new one, and ensures LocalState persists/restores the aliased model identifier consistently.

Changes

Cohort / File(s)	Summary
Model configuration & aliasing frontend/src/components/ModelSelector.tsx, frontend/src/utils/utils.ts, frontend/src/state/LocalStateContext.tsx	Added public model kimi-k2-5 (displayName/shortName "Kimi K2.5", badges ["Pro","New"], requiresPro, supportsVision, tokenLimit 256000); removed qwen3-coder-480b; mapped math category to kimi-k2-5; added alias mapping qwen3-coder-480b → kimi-k2-5; LocalState now persists/restores the aliased model name.
Marketing labels frontend/src/components/Marketing.tsx	Updated AI_MODELS labels: Kimi Moonshot entry includes "Kimi K2.5" alongside "Kimi K2"; Qwen entry consolidated to single label "Qwen3-VL".

Sequence Diagram(s)

(omitted)

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• #214: Related removal/aliasing of qwen3-coder-480b and addition of kimi-k2-5 in model configuration.
• #157: Similar changes touching MODEL_CONFIG, aliasing, and LocalState model handling.
• #196: Modifies LocalStateContext.tsx to persist/restore aliased model names.

Poem

🐰 I hopped through code with gentle paws,
Old Qwen trimmed, Kimi earns applause,
Aliases stitched, state held tight,
New badges gleam in morning light—
Hop, celebrate the model swap tonight! 🎉

🚥 Pre-merge checks | ✅ 2

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'chore: replace Qwen3 Coder with Kimi K2.5' directly and clearly describes the main change: replacing the Qwen3 Coder model with Kimi K2.5 across the codebase.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch chore/replace-qwen3-coder-with-kimi-k2-5

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[cloudflare-workers-and-pages]

Deploying maple with Cloudflare Pages

Latest commit:	055d94c
Status:	✅ Deploy successful!
Preview URL:	https://32287330.maple-ca8.pages.dev
Branch Preview URL:	https://chore-replace-qwen3-coder-wi.maple-ca8.pages.dev

View logs

[devin-ai-integration]

✅ Devin Review: No Issues Found

Devin Review analyzed this PR and found no potential bugs to report.

View in Devin Review to see 4 additional flags.

[greptile-apps]

Greptile Overview

Greptile Summary

This PR replaces the Qwen3 Coder 480B model with Kimi K2.5 as the default Math/Coding model. The changes ensure backward compatibility by aliasing stored qwen3-coder-480b references to kimi-k2-5.

Key Changes:

• Added new kimi-k2-5 model configuration (256k context, vision-capable, Pro tier)
• Removed qwen3-coder-480b from model configurations
• Updated Math/Coding category to use kimi-k2-5 as default
• Implemented migration alias in aliasModelName() to automatically map old model ID to new one
• Applied aliasing when loading stored model selection from localStorage
• Updated marketing badges to reflect new model availability
• Removed "New" badges from kimi-k2-thinking and qwen3-vl-30b (no longer new)

Confidence Score: 5/5

• This PR is safe to merge with no issues found
• The changes are well-structured with proper backward compatibility through aliasing. The migration pattern follows existing conventions in the codebase and ensures users with stored qwen3-coder-480b preferences seamlessly transition to kimi-k2-5
• No files require special attention

Important Files Changed

Filename	Overview
frontend/src/components/Marketing.tsx	Updated marketing badges to add Kimi K2.5 and remove Qwen3 Coder from displayed labels
frontend/src/components/ModelSelector.tsx	Added kimi-k2-5 model config, removed qwen3-coder-480b, updated math category default to kimi-k2-5, removed New badge from kimi-k2-thinking and qwen3-vl-30b
frontend/src/state/LocalStateContext.tsx	Applied aliasModelName to stored model selection on load and refactored persistChat to apply aliasing consistently
frontend/src/utils/utils.ts	Added migration alias mapping qwen3-coder-480b to kimi-k2-5 in aliasModelName function

Sequence Diagram

sequenceDiagram
participant User
participant LocalStorage
participant LocalStateContext
participant aliasModelName
participant ModelSelector
participant Marketing
Note over User,Marketing: Model Migration Flow
User->>LocalStateContext: App loads
LocalStateContext->>LocalStorage: getItem("selectedModel")
LocalStorage-->>LocalStateContext: "qwen3-coder-480b"
LocalStateContext->>aliasModelName: aliasModelName("qwen3-coder-480b")
aliasModelName-->>LocalStateContext: "kimi-k2-5"
LocalStateContext->>LocalStateContext: Set initial model to "kimi-k2-5"
Note over User,Marketing: Persisting Chats
User->>LocalStateContext: persistChat(chat)
LocalStateContext->>aliasModelName: aliasModelName(chat.model)
aliasModelName-->>LocalStateContext: Aliased model name
LocalStateContext->>LocalStorage: Save chat with aliased model
Note over User,Marketing: Model Selection UI
User->>ModelSelector: Selects Math/Coding category
ModelSelector->>ModelSelector: Uses CATEGORY_MODELS.math
ModelSelector->>LocalStateContext: setModel("kimi-k2-5")
LocalStateContext->>LocalStorage: setItem("selectedModel", "kimi-k2-5")
Note over User,Marketing: Marketing Display
User->>Marketing: Views marketing page
Marketing->>Marketing: Display "Kimi K2.5" badge
Marketing->>Marketing: Hide "Qwen3 Coder" badge

Loading