Add support for Gemini 3 models

README.md

@@ -200,12 +200,13 @@ Note, this will use more tokens as it generates more results.
 Default: `gemini-2.5-flash`
 
 The Gemini model to use. Supported models include:
+- `gemini-3.1-pro-preview` - Gemini 3.1 Pro Preview (State-of-the-art)
+- `gemini-3.1-flash-lite-preview` - Gemini 3.1 Flash Lite Preview (High-volume)
+- `gemini-3-flash-preview` - Gemini 3 Flash Preview (Fast and capable)
 - `gemini-2.5-flash` (default) - Fast and efficient
 - `gemini-2.5-pro` - More capable but slower
-- `gemini-1.5-flash` - Previous generation fast model
-- `gemini-1.5-pro` - Previous generation capable model
 
-> Tip: `gemini-2.5-pro` provides more sophisticated code analysis but may be slower and more expensive than `gemini-2.5-flash`.
+> Tip: `gemini-3.1-pro-preview` provides state-of-the-art reasoning but may be slower than Flash models.
 
 #### timeout
 

src/utils/config.ts

@@ -86,10 +86,11 @@ const configParsers = {
 
 		// Validate that it's a supported Gemini model
 		const supportedModels = [
+			'gemini-3.1-pro-preview',
+			'gemini-3.1-flash-lite-preview',
+			'gemini-3-flash-preview',
 			'gemini-2.5-flash',
-			'gemini-2.5-pro',
-			'gemini-1.5-flash',
-			'gemini-1.5-pro'
+			'gemini-2.5-pro'
 		];
 
 		if (!supportedModels.includes(model)) {

src/utils/gemini.ts

@@ -51,10 +51,15 @@
 					]
 				}
 			],
 			config: {
 				...generationConfig,
 				systemInstruction,
 				...(model.includes('2.5') && model.toLowerCase().includes('flash') ? { thinkingConfig: { thinkingBudget: 0 } } : {}),
+				...(model.startsWith('gemini-3') ? {
+					thinkingConfig: {
+						thinkingLevel: model.includes('pro') ? 'low' : 'minimal'
+					}
+				} : {}),
 			}
 		});