fix: gemini minimal reasoning effort

[narengogi]

Description: (required)

• Gemini supports new reasoning modes (minimal and medium) (previously only low and high were supported) for the gemini 3 flash models, this PR just updates the mapping

payload to test

{
    "model": "gemini-3-flash-preview",
    "stream": false,
    "stop": ["NO_RESPONSE_REQUIRED"],
    "reasoning_effort": "minimal",
    "messages": [
        {
            "role": "user",
            "content": [
                {
                    "type": "text",
                    "text": "Give me the flight ticket details from BLR to LDX"
                }
            ]
        }
    ]
}

Type of Change:

• Refactoring (no functional changes)

[Copilot]

Pull request overview

This PR updates Gemini's reasoning effort mapping to support the newly introduced "minimal" and "medium" modes for Gemini 3 Flash models. Previously, the code mapped OpenAI-style reasoning efforts to Vertex AI's thinking levels (consolidating minimal/low to "low" and medium/high to "high"), but now passes the reasoning_effort value directly without transformation.

Key changes:

• Removed the mapping function that converted reasoning effort values to thinking levels
• Updated both Google and Google Vertex AI providers to use reasoning_effort values directly

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
src/providers/google/chatComplete.ts	Removed mapping function import and usage; now passes reasoning_effort directly to thinkingConfig
src/providers/google-vertex-ai/transformGenerationConfig.ts	Removed the openaiReasoningEffortToVertexThinkingLevel mapping function and updated to use reasoning_effort directly

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 2 out of 2 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[davidwhite-spec]

@narengogi @VisargD hi.

We were using gemini 3 flash with reasoning effort set to "low" and we saw our latency drop like crazy from 11s to 3s. Can you help me understand what is going on? Was I truly using "low" before? Am I truly using "low" now?

[narengogi]

@davidwhite-spec haha, yes there was no minimal and medium setting previously in the gemini spec, it was added for the gemini 3 flash models, so previously to avoid errors minimal and low were mapped to low, now they are not