feat(providers): add Gemini 3.6 Flash and 3.5 Flash-Lite

[waleedlatif1]

Summary

• Add gemini-3.6-flash to the Google provider registry (input $1.50/M, cached $0.15/M, output $7.50/M) and set it as the recommended Google flash model
• Add gemini-3.5-flash-lite (input $0.30/M, cached $0.03/M, output $2.50/M), marked speed-optimized
• Move recommended from gemini-3.5-flash to gemini-3.6-flash (direct successor: better performance, 17% fewer output tokens)
• Pricing verified against Google's live API pricing docs; context window (1,048,576) and max output (65,536) match the rest of the Gemini 3.x flash family
• Excluded Gemini 3.5 Flash Cyber — limited-access via CodeMender, not a general API model

Type of Change

• New feature

Testing

Tested manually — bun run lint passes; both models fall under the google provider so they're automatically hosted + billed.

Checklist

• Code follows project style guidelines
• Self-reviewed my changes
• Tests added/updated and passing
• No new warnings introduced
• I confirm that I have read and agree to the terms outlined in the Contributor License Agreement (CLA)

[vercel]

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

1 Skipped Deployment

Project	Deployment	Actions	Updated (UTC)
docs	Skipped		Jul 21, 2026 8:09pm

[cursor]

PR Summary

Low Risk
Static provider metadata only; affects model pickers, pricing display, and recommendations—not auth or execution paths.

Overview
Registers gemini-3.6-flash and gemini-3.5-flash-lite in the Google provider catalog in models.ts, with pricing, thinking levels, 1M context, and 65k max output aligned with other Gemini 3.x flash models.

gemini-3.6-flash is marked recommended (successor to 3.5 flash; output pricing $7.50/M vs $9.00/M on gemini-3.5-flash). gemini-3.5-flash-lite is marked speedOptimized with lower per-token rates. The recommended flag is removed from gemini-3.5-flash so defaults and suggestion paths favor 3.6 flash.

No provider routing or API code changes—new IDs are picked up via existing ^gemini/ patterns and hosted Google billing.

^{Reviewed by Cursor Bugbot for commit 994d334. Configure here.}

[greptile-apps]

Greptile Summary

This PR expands the Google model registry with two Gemini Flash models.

• Adds Gemini 3.6 Flash with pricing and capability metadata.
• Adds the speed-optimized Gemini 3.5 Flash-Lite model.
• Makes Gemini 3.6 Flash the recommended Google Flash model.

Confidence Score: 5/5

This looks safe to merge.

• The generated capability collections include both new models from their registry metadata.
• No blocking issue remains in the changed code.

Important Files Changed

Filename	Overview
apps/sim/providers/models.ts	Adds two Gemini model definitions and moves the recommended designation to Gemini 3.6 Flash.

_{Reviews (2): Last reviewed commit: "feat(providers): add Gemini 3.6 Flash an..." | Re-trigger Greptile}

[waleedlatif1]

@greptile the one flagged thread was a false positive — the temperature/thinking support lists are derived at module load from each model entry's capabilities (getModelsWithTemperatureSupport/getModelsWithThinking iterate PROVIDER_DEFINITIONS), so both new IDs are included automatically. Verified with temporary assertions in providers/utils.test.ts (152/152 pass). Please re-review.

[waleedlatif1]

@cursor review

[cursor]

✅ Bugbot reviewed your changes and found no new issues!

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 994d334. Configure here.}

[waleedlatif1]

Verified both models against the live Gemini API (generativelanguage.googleapis.com/v1beta):

Field	Entry	Live API
gemini-3.6-flash contextWindow	1048576	inputTokenLimit: 1048576 ✅
gemini-3.6-flash maxOutputTokens	65536	outputTokenLimit: 65536 ✅
gemini-3.5-flash-lite contextWindow	1048576	inputTokenLimit: 1048576 ✅
gemini-3.5-flash-lite maxOutputTokens	65536	outputTokenLimit: 65536 ✅
temperature max (both)	2	maxTemperature: 2 ✅
thinking (both)	enabled	thinking: true ✅
cached input pricing (both)	set	createCachedContent supported ✅

Both models return HTTP 200 on generateContent with serviceTier: standard (matches the Standard-tier pricing used). Pricing ($1.50/$0.15/$7.50 and $0.30/$0.03/$2.50 per 1M) two-source verified (Google pricing docs + OpenRouter); release date 2026-07-21 confirmed by OpenRouter.