From cd0366392ae20920ec6590e700d34d57b4a46dc0 Mon Sep 17 00:00:00 2001
From: AC <ac@fabtek.cn>
Date: Wed, 11 Jun 2025 15:25:38 +0800
Subject: [PATCH] change max tokens

---
 app/api/bedrock/index.ts        | 2 +-
 app/client/platforms/bedrock.ts | 2 +-
 app/client/platforms/openai.ts  | 2 +-
 app/store/chat.ts               | 2 +-
 app/store/config.ts             | 2 +-
 5 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/app/api/bedrock/index.ts b/app/api/bedrock/index.ts
index a35cfc55a..112fa9c38 100644
--- a/app/api/bedrock/index.ts
+++ b/app/api/bedrock/index.ts
@@ -374,7 +374,7 @@ export async function handle(
     const payload = {
       anthropic_version: "bedrock-2023-05-31",
       max_tokens:
-        typeof max_tokens === "number" && max_tokens > 0 ? max_tokens : 4096,
+        typeof max_tokens === "number" && max_tokens > 0 ? max_tokens : 8000,
       temperature:
         typeof temperature === "number" && temperature >= 0 && temperature <= 1
           ? temperature
diff --git a/app/client/platforms/bedrock.ts b/app/client/platforms/bedrock.ts
index 2e0df43b8..b5f12c542 100644
--- a/app/client/platforms/bedrock.ts
+++ b/app/client/platforms/bedrock.ts
@@ -31,7 +31,7 @@ export class BedrockApi implements LLMApi {
           messages,
           temperature: modelConfig.temperature,
           stream: !!modelConfig.stream,
-          max_tokens: (modelConfig as any).max_tokens || 4096, // Cast to access max_tokens from ModelConfig
+          max_tokens: (modelConfig as any).max_tokens || 8000, // Cast to access max_tokens from ModelConfig
         }),
         signal: controller.signal,
         headers: getHeaders(), // getHeaders should handle Bedrock (no auth needed)
diff --git a/app/client/platforms/openai.ts b/app/client/platforms/openai.ts
index c6f3fc425..350ea272c 100644
--- a/app/client/platforms/openai.ts
+++ b/app/client/platforms/openai.ts
@@ -244,7 +244,7 @@ export class ChatGPTApi implements LLMApi {
 
       // add max_tokens to vision model
       if (visionModel) {
-        requestPayload["max_tokens"] = Math.max(modelConfig.max_tokens, 4000);
+        requestPayload["max_tokens"] = Math.max(modelConfig.max_tokens, 8000);
       }
     }
 
diff --git a/app/store/chat.ts b/app/store/chat.ts
index 6c98923da..42c32543f 100644
--- a/app/store/chat.ts
+++ b/app/store/chat.ts
@@ -867,7 +867,7 @@ export const useChatStore = createPersistStore(
 
         const historyMsgLength = countMessages(toBeSummarizedMsgs);
 
-        if (historyMsgLength > (modelConfig?.max_tokens || 4000)) {
+        if (historyMsgLength > (modelConfig?.max_tokens || 8000)) {
           const n = toBeSummarizedMsgs.length;
           toBeSummarizedMsgs = toBeSummarizedMsgs.slice(
             Math.max(0, n - modelConfig.historyMessageCount),
diff --git a/app/store/config.ts b/app/store/config.ts
index 45e21b026..eefb39f41 100644
--- a/app/store/config.ts
+++ b/app/store/config.ts
@@ -68,7 +68,7 @@ export const DEFAULT_CONFIG = {
     providerName: "OpenAI" as ServiceProvider,
     temperature: 0.5,
     top_p: 1,
-    max_tokens: 4000,
+    max_tokens: 8000,
     presence_penalty: 0,
     frequency_penalty: 0,
     sendMemory: true,