diff --git a/packages/ai/src/provider-error.ts b/packages/ai/src/provider-error.ts
index 5f6574846e8..35b9cff7425 100644
--- a/packages/ai/src/provider-error.ts
+++ b/packages/ai/src/provider-error.ts
@@ -21,6 +21,8 @@ const patterns = [
   /exceeds the context window/i,
   /exceeds (?:the )?(?:model'?s )?maximum context length(?: of [\d,]+ tokens?|\s*\([\d,]+\))/i,
   /input token count.*exceeds the maximum/i,
+  // Amazon Nova on Bedrock reports this as a mid-stream validationException.
+  /number of input tokens exceeds maximum length/i,
   /tokens in request more than max tokens allowed/i,
   /maximum prompt length is \d+/i,
   /reduce the length of the messages/i,
diff --git a/packages/ai/test/provider-error.test.ts b/packages/ai/test/provider-error.test.ts
index 52599f8ae3f..98a94c1b143 100644
--- a/packages/ai/test/provider-error.test.ts
+++ b/packages/ai/test/provider-error.test.ts
@@ -366,6 +366,23 @@ describe("provider error rawBody classification", () => {
     expect(reason._tag === "InvalidRequest" ? reason.classification : reason._tag).toBe("context-overflow")
   })
 
+  test("classifies Amazon Nova input token rejections as context overflow", () => {
+    const message =
+      "The model returned the following errors: Input Tokens Exceeded: Number of input tokens exceeds maximum length. Please update the input to try again."
+    const reason = classifyProviderFailure({
+      message,
+      rawBody: JSON.stringify({
+        headers: {
+          ":exception-type": { type: "string", value: "validationException" },
+          ":content-type": { type: "string", value: "application/json" },
+          ":message-type": { type: "string", value: "exception" },
+        },
+        body: JSON.stringify({ message }),
+      }),
+    })
+    expect(reason._tag === "InvalidRequest" ? reason.classification : reason._tag).toBe("context-overflow")
+  })
+
   test("classifies Google invalid API keys as authentication failures", () => {
     const rawBody = JSON.stringify({
       error: {