fix(openrouter): conditionally include max_tokens parameter

- Set default maxTokens to 0 in getOpenRouterModels
- Only include max_tokens in API requests when value > 0
- Preserves OpenRouter defaults when no override exists

Author: daniel-lxs

src/api/providers/fetchers/openrouter.ts

@@ -72,7 +72,7 @@ export async function getOpenRouterModels(options?: ApiHandlerOptions): Promise<
 				typeof cacheWritesPrice !== "undefined" && typeof cacheReadsPrice !== "undefined"
 
 			const modelInfo: ModelInfo = {
-				maxTokens: rawModel.top_provider?.max_completion_tokens,
+				maxTokens: 0,
 				contextWindow: rawModel.context_length,
 				supportsImages: rawModel.architecture?.modality?.includes("image"),
 				supportsPromptCache,

src/api/providers/openrouter.ts

@@ -106,7 +106,7 @@ export class OpenRouterHandler extends BaseProvider implements SingleCompletionH
 		// https://openrouter.ai/docs/transforms
 		const completionParams: OpenRouterChatCompletionParams = {
 			model: modelId,
-			max_tokens: maxTokens,
+			...(maxTokens && maxTokens > 0 && { max_tokens: maxTokens }),
 			temperature,
 			thinking, // OpenRouter is temporarily supporting this.
 			top_p: topP,