Replace retired models with Gemma 4 and Kimi K2.5

[AnthonyRonning]

Summary

• add Gemma 4 31B to the active lineup with Starter access, vision support, and updated selector badges
• retire Gemma 3 and DeepSeek from the app, pricing copy, and public model docs while migrating legacy selections onto supported models
• update pricing, upsell, marketing, and FAQ copy to reflect the current model lineup and legacy Starter behavior

Testing

• just format
• just lint
• just build
• bun test

---

Summary by CodeRabbit

• New Features

  • Added Gemma 4 31B and Kimi K2.5 to the model lineup; GPT-OSS 120B now listed on Free tier alongside Llama 3.3 70B

• Bug Fixes

  • Improved context-limit guidance by removing model-specific prompts and simplifying recommendations

• Documentation

  • Updated AI model listings, availability, and pricing language across marketing, pricing, and help docs; removed DeepSeek and older Gemma 3 references

• UI

  • Model badges and selector now reflect the updated model lineup and vision-capable defaults

[cloudflare-workers-and-pages]

Deploying maple with    Cloudflare Pages

Latest commit:	1dc15f0
Status:	✅  Deploy successful!
Preview URL:	https://53e63d58.maple-ca8.pages.dev
Branch Preview URL:	https://add-gemma4-31b-model.maple-ca8.pages.dev

View logs

[coderabbitai]

📝 Walkthrough

Walkthrough

This PR updates the AI model lineup across docs, UI, pricing, and model-selection logic: removes DeepSeek R1 and Gemma 3 27B, adds Gemma 4 31B and Kimi K2.5, adjusts image/vision auto-selection, and updates related copy and aliases.

Changes

Cohort / File(s)	Summary
Documentation docs/support-faq.md, frontend/public/llms-full.txt, frontend/public/llms.txt	Replaced DeepSeek R1 and Gemma 3 references with Gemma 4 31B and Kimi K2.5; updated model availability, context-window and image-capability descriptions, and developer example model id.
Model config & aliasing frontend/src/components/ModelSelector.tsx, frontend/src/utils/utils.ts	Removed legacy model IDs (gemma-3-27b, leon-se/gemma-3-27b-it-fp8-dynamic, deepseek-r1-0528); added gemma4-31b (256K, vision) and badge for gpt-oss-120b; added aliases mapping old IDs to new (gemma4-31b, kimi-k2-5); changed image auto-switch to prefer gemma4-31b.
UI badges & copy frontend/src/components/Marketing.tsx, frontend/src/routes/teams.tsx, frontend/src/components/PromoDialog.tsx	Removed DeepSeek badge, updated Google badge labels to “Gemma 4”, added Kimi K2.5 entry; updated promo/marketing copy to reference Gemma 4 and Kimi K2.5.
Dialogs & pricing copy frontend/src/components/ContextLimitDialog.tsx, frontend/src/components/UpgradePromptDialog.tsx, frontend/src/config/pricingConfig.tsx	Removed Gemma-specific context guidance; updated upgrade and usage messaging to highlight Gemma 4, Qwen3-VL, and Kimi K2.5; updated pricing plan feature text to reflect new lineup.

Sequence Diagram(s)

(omitted)

Estimated code review effort

🎯 3 (Moderate) | ⏱️ ~20 minutes

Possibly related PRs

• Add support for deepseek-r1-0528 model #168 — modifies MODEL_CONFIG and pricing/model config entries; likely overlaps on removing/adding model IDs and badge/availability updates.
• Add Kimi K2 Thinking model and re-enable brain toggle #346 — updates MODEL_CONFIG, Kimi/DeepSeek mappings, and model-selection behavior; closely related to aliasing and model replacements here.
• Remove kimi-k2 model and brain toggle, keep kimi-k2.5 #414 — touches marketing, ModelSelector, and alias utilities to standardize Kimi/DeepSeek naming and badge changes; strong overlap in affected areas.

Poem

🐰 Models hop and badges rearrange,
DeepSeek leaves; new lights take the range.
Gemma 4 bounds with vision bright,
Kimi K2.5 leaps into flight,
Configs updated — carrots celebrated! 🥕

🚥 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 PR title directly and clearly summarizes the main change: retiring outdated models (Gemma 3 and DeepSeek) and replacing them with newer ones (Gemma 4 and Kimi K2.5), which is the core objective across all modified files.

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

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch add-gemma4-31b-model

---

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

[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 5 additional findings.

[AnthonyRonning]

@TestFlight build

[github-actions]

🚀 TestFlight deployment triggered! Check the Actions tab for progress.

[github-actions]

✅ TestFlight deployment completed successfully!