diff --git a/src/chat.js b/src/chat.js
index 6626202..bd815c1 100644
--- a/src/chat.js
+++ b/src/chat.js
@@ -2,7 +2,7 @@
 import { marked } from "https://cdn.jsdelivr.net/npm/marked/lib/marked.esm.js";
 import hljs from "https://cdnjs.cloudflare.com/ajax/libs/highlight.js/11.8.0/es/highlight.min.js";
 import { API_KEY, getAPIConfig } from "./config.js";
-import { getSelectedModel, getSelectedModelEndpoint, getSelectedModelMaxTokens } from "./models.js";
+import { getSelectedModel, getSelectedModelEndpoint, getSelectedModelMaxTokens, getSelectedModelMaxCompletionTokens } from "./models.js";
 
 // Rough token estimation (4 chars = 1 token average for English)
 function estimateTokens(text) {
@@ -93,7 +93,11 @@ export async function* sendMessage(message) {
 
 	const modelMaxTokens = getSelectedModelMaxTokens();
 	const calculatedAvailable = calculateAvailableTokens(chatHistory, modelMaxTokens);
-	console.log("Chat using calculated available tokens:", calculatedAvailable, "of model max:", modelMaxTokens);
+	
+	// Cap to model's max completion tokens limit
+	const maxCompletionTokens = getSelectedModelMaxCompletionTokens();
+	const finalAvailable = Math.min(calculatedAvailable, maxCompletionTokens);
+	console.log("Chat using tokens:", finalAvailable, "calculated:", calculatedAvailable, "max completion:", maxCompletionTokens);
 	
 	const response = await fetch(apiUrl, {
 		method: "POST",
@@ -102,7 +106,7 @@ export async function* sendMessage(message) {
 			model: model,
 			messages: chatHistory,
 			temperature: 0.5,
-			max_tokens: calculatedAvailable,
+			max_tokens: finalAvailable,
 			stream: true,
 		}),
 	});
@@ -256,8 +260,12 @@ export async function* sendMessageWithCustomHistory(customHistory) {
 	}
 
 	const modelMaxTokens = getSelectedModelMaxTokens();
-	const availableTokens = calculateAvailableTokens(customHistory, modelMaxTokens);
-	console.log("Chat (custom history) using max_tokens:", availableTokens);
+	const calculatedAvailable = calculateAvailableTokens(customHistory, modelMaxTokens);
+	
+	// Cap to model's max completion tokens limit
+	const maxCompletionTokens = getSelectedModelMaxCompletionTokens();
+	const finalAvailable = Math.min(calculatedAvailable, maxCompletionTokens);
+	console.log("Chat (custom history) using tokens:", finalAvailable, "calculated:", calculatedAvailable, "max completion:", maxCompletionTokens);
 	
 	const response = await fetch(apiUrl, {
 		method: "POST",
@@ -265,7 +273,7 @@ export async function* sendMessageWithCustomHistory(customHistory) {
 		body: JSON.stringify({
 			model: model,
 			messages: customHistory,
-			max_tokens: availableTokens,
+			max_tokens: finalAvailable,
 			stream: true,
 		}),
 	});
diff --git a/src/tts.js b/src/tts.js
index e078d15..e3e7c51 100644
--- a/src/tts.js
+++ b/src/tts.js
@@ -145,7 +145,11 @@ Your task:
 Return only the article text ready for TTS, nothing else.`;
 	const inputTokens = Math.ceil((systemPrompt + preCleanedContent).length / 4);
 	const buffer = Math.max(2048, Math.floor(inputTokens * 1.5)); // Use 1.5x input tokens as buffer, minimum 2048
-	const availableTokens = Math.max(100, modelMaxTokens - inputTokens - buffer);
+	const calculatedAvailable = Math.max(100, modelMaxTokens - inputTokens - buffer);
+	
+	// Cap to model's max completion tokens limit
+	const maxCompletionTokens = getSelectedModelMaxCompletionTokens();
+	const availableTokens = Math.min(calculatedAvailable, maxCompletionTokens);
 
 	const payload = {
 		model: getSelectedModel(),
diff --git a/src/ui.js b/src/ui.js
index 3c6314f..61eeabb 100644
--- a/src/ui.js
+++ b/src/ui.js
@@ -138,7 +138,7 @@ async function* sendMessageWithCustomHistory(messageHistory) {
 	}
 
 	// Import getSelectedModelMaxTokens to get dynamic max tokens and calculate available
-	const { getSelectedModelMaxTokens } = await import("./models.js");
+	const { getSelectedModelMaxTokens, getSelectedModelMaxCompletionTokens } = await import("./models.js");
 	const modelMaxTokens = getSelectedModelMaxTokens();
 	
 	// Estimate input tokens for intro generation
@@ -146,7 +146,11 @@ async function* sendMessageWithCustomHistory(messageHistory) {
 	const inputTokens = Math.ceil(inputText.length / 4);
 	const buffer = Math.max(2048, Math.floor(inputTokens * 1.5)); // Use 1.5x input tokens as buffer, minimum 2048
 	const calculatedAvailable = Math.max(100, modelMaxTokens - inputTokens - buffer);
-	console.log("UI (intro generation) using calculated available tokens:", calculatedAvailable, "buffer:", buffer, "of model max:", modelMaxTokens);
+	
+	// Cap to model's max completion tokens limit
+	const maxCompletionTokens = getSelectedModelMaxCompletionTokens();
+	const finalAvailable = Math.min(calculatedAvailable, maxCompletionTokens);
+	console.log("UI (intro generation) using tokens:", finalAvailable, "calculated:", calculatedAvailable, "max completion:", maxCompletionTokens, "of model max:", modelMaxTokens);
 	
 	const response = await fetch(apiUrl, {
 		method: "POST",
@@ -155,7 +159,7 @@ async function* sendMessageWithCustomHistory(messageHistory) {
 			model: model,
 			messages: messageHistory,
 			temperature: 0.3,
-			max_tokens: calculatedAvailable,
+			max_tokens: finalAvailable,
 			stream: true,
 		}),
 	});