Adds GPT-5 models and enforces temperature reqs

 - Ensures unset temperature for GPT5 models as it is unsupported

Author: eamodio

CHANGELOG.md

@@ -16,6 +16,8 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/) and this p
 - Adds ability to set or change the upstream branch for branches in the _Commit Graph_ and other GitLens views ([#4498](https://github.com/gitkraken/vscode-gitlens/issues/4498))
   - Adds new _Set Upstream..._ and _Change Upstream..._ context menu items to branches in the _Commit Graph_ and other GitLens views
   - Adds a new _upstream_ sub-command to the _branch_ Git Command Palette
+- Adds updated AI model support for GitLens' AI features
+  - Adds GPT-5 family (GPT-5, GPT-5 Mini, GPT-5 Nano), and Claude 4.1 Opus models
 - Add Azure DevOps Server integration support ([#4478](https://github.com/gitkraken/vscode-gitlens/issues/4478))
 
 ### Changed

src/plus/ai/models/model.ts

@@ -97,6 +97,45 @@ export const openAIModels = <T extends OpenAIProviders>(provider: AIProviderDesc
 		provider: provider,
 		hidden: true,
 	},
+	{
+		id: 'gpt-5',
+		name: 'GPT-5',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+	},
+	{
+		id: 'gpt-5-2025-08-07',
+		name: 'GPT-5',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+		hidden: true,
+	},
+	{
+		id: 'gpt-5-mini',
+		name: 'GPT-5 mini',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+	},
+	{
+		id: 'gpt-5-mini-2025-08-07',
+		name: 'GPT-5 mini',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+		hidden: true,
+	},
+	{
+		id: 'gpt-5-nano',
+		name: 'GPT-5 nano',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+	},
+	{
+		id: 'gpt-5-nano-2025-08-07',
+		name: 'GPT-5 nano',
+		maxTokens: { input: 400000, output: 128000 },
+		provider: provider,
+		hidden: true,
+	},
 	{
 		id: 'o4-mini',
 		name: 'o4 mini',

src/plus/ai/openAICompatibleProviderBase.ts

@@ -130,7 +130,7 @@ export abstract class OpenAICompatibleProviderBase<T extends AIProviders> implem
 				max_completion_tokens: model.maxTokens.output
 					? Math.min(modelOptions?.outputTokens ?? Infinity, model.maxTokens.output)
 					: modelOptions?.outputTokens,
-				temperature: getValidatedTemperature(modelOptions?.temperature ?? model.temperature),
+				temperature: getValidatedTemperature(model, modelOptions?.temperature ?? model.temperature),
 			};
 
 			const rsp = await this.fetchCore(action, model, apiKey, request, cancellation);

src/plus/ai/utils/-webview/ai.utils.ts

@@ -137,10 +137,13 @@ export async function getOrPromptApiKey(
 	return apiKey;
 }
 
-export function getValidatedTemperature(modelTemperature?: number | null): number | undefined {
+export function getValidatedTemperature(model: AIModel, modelTemperature?: number | null): number | undefined {
 	if (modelTemperature === null) return undefined;
-	if (modelTemperature != null) return modelTemperature;
-	return Math.max(0, Math.min(configuration.get('ai.modelOptions.temperature'), 2));
+	// GPT5 doesn't support anything but the default temperature
+	if (model.id.startsWith('gpt-5')) return undefined;
+
+	modelTemperature ??= Math.max(0, Math.min(configuration.get('ai.modelOptions.temperature'), 2));
+	return modelTemperature;
 }
 
 export async function showLargePromptWarning(estimatedTokens: number, threshold: number): Promise<boolean> {

src/plus/ai/vscodeProvider.ts

@@ -100,7 +100,7 @@ export class VSCodeAIProvider implements AIProvider<typeof provider.id> {
 							outputTokens: model.maxTokens.output
 								? Math.min(options.modelOptions?.outputTokens ?? Infinity, model.maxTokens.output)
 								: options.modelOptions?.outputTokens,
-							temperature: getValidatedTemperature(model.temperature),
+							temperature: getValidatedTemperature(model, model.temperature),
 						},
 					},
 					options.cancellation,