From e28255edb6f0955332e827bb706bbcd6465d1ff6 Mon Sep 17 00:00:00 2001
From: Mingholy <mingholy.lmh@gmail.com>
Date: Tue, 14 Oct 2025 15:40:20 +0800
Subject: [PATCH] fix: token limits for qwen3-max (#724)

---
 .vscode/launch.json                   | 11 +++++++++++
 packages/core/src/core/tokenLimits.ts |  6 +++---
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/.vscode/launch.json b/.vscode/launch.json
index 8968c578..1966371c 100644
--- a/.vscode/launch.json
+++ b/.vscode/launch.json
@@ -108,6 +108,17 @@
       "request": "attach",
       "skipFiles": ["<node_internals>/**"],
       "type": "node"
+    },
+    {
+      "type": "node",
+      "request": "launch",
+      "name": "Debug Current TS File",
+      "runtimeExecutable": "npx",
+      "runtimeArgs": ["tsx", "${file}"],
+      "skipFiles": ["<node_internals>/**"],
+      "cwd": "${workspaceFolder}",
+      "console": "integratedTerminal",
+      "env": {}
     }
   ],
   "inputs": [
diff --git a/packages/core/src/core/tokenLimits.ts b/packages/core/src/core/tokenLimits.ts
index f8ed1018..5992e8f0 100644
--- a/packages/core/src/core/tokenLimits.ts
+++ b/packages/core/src/core/tokenLimits.ts
@@ -115,7 +115,7 @@ const PATTERNS: Array<[RegExp, TokenCount]> = [
   [/^coder-model$/, LIMITS['1m']],
 
   // Commercial Qwen3-Max-Preview: 256K token context
-  [/^qwen3-max-preview(-.*)?$/, LIMITS['256k']], // catches "qwen3-max-preview" and date variants
+  [/^qwen3-max(-preview)?(-.*)?$/, LIMITS['256k']], // catches "qwen3-max" or "qwen3-max-preview" and date variants
 
   // Open-source Qwen3-Coder variants: 256K native
   [/^qwen3-coder-.*$/, LIMITS['256k']],
@@ -179,8 +179,8 @@ const OUTPUT_PATTERNS: Array<[RegExp, TokenCount]> = [
   // Generic coder-model: same as qwen3-coder-plus (64K max output tokens)
   [/^coder-model$/, LIMITS['64k']],
 
-  // Qwen3-Max-Preview: 65,536 max output tokens
-  [/^qwen3-max-preview(-.*)?$/, LIMITS['64k']],
+  // Qwen3-Max: 65,536 max output tokens
+  [/^qwen3-max(-preview)?(-.*)?$/, LIMITS['64k']],
 
   // Qwen-VL-Max-Latest: 8,192 max output tokens
   [/^qwen-vl-max-latest$/, LIMITS['8k']],