diff --git a/packages/ai/CHANGELOG.md b/packages/ai/CHANGELOG.md
index 74aa15c4d..30b17a788 100644
--- a/packages/ai/CHANGELOG.md
+++ b/packages/ai/CHANGELOG.md
@@ -5,6 +5,7 @@
 ### Fixed
 
 - Fixed Claude models that accept mid-conversation system messages rejecting a session with a 400 (`role 'system' must precede an 'assistant' message`) when a system update was followed by an assistant reply with nothing to send to Anthropic, such as an empty reply or another model's reasoning-only reply ([#11091](https://github.com/earendil-works/pi/issues/11091))
+- Fixed Claude models with mid-conversation effort failing with `Invalid effort level` on accounts whose organization caps effort below `high`, even at `low` or `medium` thinking, by sending a fixed `low` top-level effort ([#10063](https://github.com/earendil-works/pi/issues/10063))
 
 ## [1.4.0] - 2026-10-21
 
diff --git a/packages/ai/src/api/anthropic-messages.ts b/packages/ai/src/api/anthropic-messages.ts
index b43b11c0e..00d820341 100644
--- a/packages/ai/src/api/anthropic-messages.ts
+++ b/packages/ai/src/api/anthropic-messages.ts
@@ -1300,7 +1300,10 @@ function buildParams(
 			display: options?.thinkingDisplay ?? "summarized",
 			block_binding: { prefix_mismatch_behavior: "drop_block" },
 		};
-		params.output_config = { effort: "high" };
+		// Effort markers set the level for every reply. The top-level effort is part of the cached
+		// prompt, so it stays fixed. It is "low" because organizations can cap effort and reject
+		// requests containing any level above the cap.
+		params.output_config = { effort: "low" };
 	} else if (model.reasoning && options?.thinkingEnabled) {
 		// Default to "summarized" so Opus 4.7 and Mythos Preview behave like
 		// older Claude 4 models (whose API default is also "summarized").
diff --git a/packages/ai/test/anthropic-mid-conversation-effort.test.ts b/packages/ai/test/anthropic-mid-conversation-effort.test.ts
index 3ba1fb2bc..54dad1d95 100644
--- a/packages/ai/test/anthropic-mid-conversation-effort.test.ts
+++ b/packages/ai/test/anthropic-mid-conversation-effort.test.ts
@@ -108,8 +108,8 @@ describe("Anthropic mid-conversation effort", () => {
 			content: [],
 			output_config: { effort: "high" },
 		});
-		expect(first.payload.output_config).toEqual({ effort: "high" });
-		expect(second.payload.output_config).toEqual({ effort: "high" });
+		expect(first.payload.output_config).toEqual({ effort: "low" });
+		expect(second.payload.output_config).toEqual({ effort: "low" });
 		expect(second.payload.thinking).toEqual({
 			type: "adaptive",
 			display: "summarized",