[BUG] Gemini 3 has incorrect pricing

[rossdonald]

Problem (one or two sentences)

The pricing in Roo does not match the pricing on the official Google API page for Gemini 3 flash.

https://ai.google.dev/gemini-api/docs/pricing#gemini-3-flash-preview

Current official pricing:

• Input price $0.50
• Output price (including thinking tokens) $3.00
• Context cache read price $0.05
• Context cache write price $0.00

Roo-Code/packages/types/src/providers/gemini.ts

Lines 50 to 53 in 586cf47

	inputPrice: 0.3,
	outputPrice: 2.5,
	cacheReadsPrice: 0.075,
	cacheWritesPrice: 1.0,

Gemini 3 Pro preview doesn't have a price for cached tokens.

Add caching price (cacheReadsPrice)
Current official context caching price:

• $0.20 for prompts <= 200k tokens
• $0.40 for prompts > 200k tokens

Roo-Code/packages/types/src/providers/gemini.ts

Lines 24 to 34 in 586cf47

	tiers: [
	{
	contextWindow: 200_000,
	inputPrice: 2.0,
	outputPrice: 12.0,
	},
	{
	contextWindow: Infinity,
	inputPrice: 4.0,
	outputPrice: 18.0,
	},

The vertex provider also has the same pricing errors for Flash and Pro.

Roo-Code/packages/types/src/providers/vertex.ts

Line 50 in 586cf47

inputPrice: 0.3,

Context (who is affected and when)

Users using the Gemini or Vertex API directly.

Reproduction steps

Use the Gemini provider or the Vertex provider and make API calls, then check the Google console pricing to see the differences.

Expected result

Price is very similar to what is shown in Google Console.

Actual result

Pricing is different.

Variations tried (optional)

No response

App Version

3.38.2

API Provider (optional)

Google Gemini

Model Used (optional)

gemini-3-flash-preview

Roo Code Task Links (optional)

No response

Relevant logs or errors (optional)

[roomote]

I have analyzed this issue and here is my proposed implementation plan:

Summary of Changes

Based on the official Google pricing documentation, I will update the pricing for Gemini 3 models in both packages/types/src/providers/gemini.ts and packages/types/src/providers/vertex.ts:

1. gemini-3-flash-preview (both files)

Current pricing:

• inputPrice: 0.3
• outputPrice: 2.5
• cacheReadsPrice: 0.075
• cacheWritesPrice: 1.0

Updated to official pricing:

• inputPrice: 0.50
• outputPrice: 3.00
• cacheReadsPrice: 0.05
• cacheWritesPrice: 0.00

2. gemini-3-pro-preview (both files)

Current pricing:

• inputPrice: 2.0 / 4.0 (tiered)
• outputPrice: 12.0 / 18.0 (tiered)
• No cache pricing

Add tiered cache pricing:

• cacheReadsPrice: 0.20 (for prompts <= 200k tokens)
• cacheReadsPrice: 0.40 (for prompts > 200k tokens)
• cacheWritesPrice: 0.00 (both tiers)

Files to Modify

• packages/types/src/providers/gemini.ts (lines 9-36, 37-54)
• packages/types/src/providers/vertex.ts (lines 9-36, 37-54)

This approach ensures consistency between the Gemini and Vertex providers and aligns with the official Google API pricing.

Please mention @roomote in a comment saying that you approve this plan, and I will proceed with the implementation.

[hannesrudolph]

@roomote here is the proper pricing

[roomote]

Thanks for the pricing details! I will update the Gemini 3 Flash and Pro pricing in both the Gemini and Vertex providers. Opening a PR shortly.

[roomote]

Opened PR #10487 with my attempt at addressing this. Updated Gemini 3 Flash and Pro pricing in both Gemini and Vertex providers. Feedback welcome!