diff --git a/open-sse/handlers/chatCore/nonStreamingHandler.js b/open-sse/handlers/chatCore/nonStreamingHandler.js
index becc0842a2a..3928ce766ff 100644
--- a/open-sse/handlers/chatCore/nonStreamingHandler.js
+++ b/open-sse/handlers/chatCore/nonStreamingHandler.js
@@ -141,17 +141,17 @@ function openAICompletionToResponses(responseBody, customToolNames = null) {
 /**
  * Translate non-streaming response body from provider format → OpenAI format.
  */
+// Provider responded in OpenAI Chat Completions shape — hand the client whatever
+// dialect it speaks, or the body unchanged when that dialect is OpenAI itself.
+function fromOpenAICompletion(responseBody, sourceFormat, customToolNames) {
+  if (sourceFormat === FORMATS.OPENAI_RESPONSES) return openAICompletionToResponses(responseBody, customToolNames);
+  if (sourceFormat === FORMATS.CLAUDE) return openAICompletionToClaudeMessage(responseBody);
+  return responseBody;
+}
+
 export function translateNonStreamingResponse(responseBody, targetFormat, sourceFormat, customToolNames = null) {
   if (targetFormat === sourceFormat) return responseBody;
-  // Provider responded in OpenAI Chat Completions shape but the client speaks
-  // Responses API — convert so tool_calls/text surface as Responses `output`.
-  if (targetFormat === FORMATS.OPENAI && sourceFormat === FORMATS.OPENAI_RESPONSES) {
-    return openAICompletionToResponses(responseBody, customToolNames);
-  }
-  if (targetFormat === FORMATS.OPENAI && sourceFormat === FORMATS.CLAUDE) {
-    return openAICompletionToClaudeMessage(responseBody);
-  }
-  if (targetFormat === FORMATS.OPENAI) return responseBody;
+  if (targetFormat === FORMATS.OPENAI) return fromOpenAICompletion(responseBody, sourceFormat, customToolNames);
 
   // Gemini / Antigravity
   if (targetFormat === FORMATS.GEMINI || targetFormat === FORMATS.ANTIGRAVITY || targetFormat === FORMATS.GEMINI_CLI || targetFormat === FORMATS.VERTEX) {
@@ -275,6 +275,10 @@ export function translateNonStreamingResponse(responseBody, targetFormat, source
     return ollamaBodyToOpenAI(responseBody);
   }
 
+  if (Array.isArray(responseBody?.choices)) {
+    return fromOpenAICompletion(responseBody, sourceFormat, customToolNames);
+  }
+
   return responseBody;
 }
 
diff --git a/tests/unit/nonstream-binary-transport-source-format-3199.test.js b/tests/unit/nonstream-binary-transport-source-format-3199.test.js
new file mode 100644
index 00000000000..4c3ac09a47e
--- /dev/null
+++ b/tests/unit/nonstream-binary-transport-source-format-3199.test.js
@@ -0,0 +1,88 @@
+import { describe, expect, it, vi } from "vitest";
+
+vi.mock("@/lib/usageDb.js", () => ({
+  appendRequestLog: vi.fn(async () => {}),
+  saveRequestDetail: vi.fn(async () => {}),
+  saveRequestUsage: vi.fn(async () => {})
+}));
+
+const { FORMATS } = await import("../../open-sse/translator/formats.js");
+const { translateNonStreamingResponse } = await import("../../open-sse/handlers/chatCore/nonStreamingHandler.js");
+
+// The kiro executor decodes its EventStream frames into a chat.completion body
+// before chatCore ever sees them, so translateNonStreamingResponse is handed
+// OpenAI shape under targetFormat "kiro".
+const KIRO_DECODED_BODY = {
+  id: "chatcmpl-1786352438097",
+  object: "chat.completion",
+  created: 1786352438,
+  model: "claude-haiku-4.5",
+  choices: [{
+    index: 0,
+    message: { role: "assistant", content: "Hey there." },
+    finish_reason: "stop",
+  }],
+  usage: { prompt_tokens: 10, completion_tokens: 8 },
+};
+
+const KIRO_TOOL_BODY = {
+  id: "chatcmpl-tool",
+  object: "chat.completion",
+  model: "claude-haiku-4.5",
+  choices: [{
+    index: 0,
+    message: {
+      role: "assistant",
+      content: null,
+      tool_calls: [{ id: "toolu_1", type: "function", function: { name: "Read", arguments: '{"path":"a.txt"}' } }],
+    },
+    finish_reason: "tool_calls",
+  }],
+  usage: {},
+};
+
+describe("#3199 non-streaming responses from binary-transport targets", () => {
+  it("returns an Anthropic message to a Claude client", () => {
+    const out = translateNonStreamingResponse(KIRO_DECODED_BODY, FORMATS.KIRO, FORMATS.CLAUDE);
+
+    expect(out.type).toBe("message");
+    expect(out.role).toBe("assistant");
+    expect(out.content).toEqual([{ type: "text", text: "Hey there." }]);
+    expect(out.stop_reason).toBe("end_turn");
+    expect(out.usage).toEqual({ input_tokens: 10, output_tokens: 8 });
+    expect(out.choices).toBeUndefined();
+  });
+
+  it("maps tool calls to tool_use blocks for a Claude client", () => {
+    const out = translateNonStreamingResponse(KIRO_TOOL_BODY, FORMATS.KIRO, FORMATS.CLAUDE);
+
+    expect(out.content).toEqual([
+      { type: "tool_use", id: "toolu_1", name: "Read", input: { path: "a.txt" } },
+    ]);
+    expect(out.stop_reason).toBe("tool_use");
+  });
+
+  it("returns a Responses body to a Responses client", () => {
+    const out = translateNonStreamingResponse(KIRO_DECODED_BODY, FORMATS.KIRO, FORMATS.OPENAI_RESPONSES);
+
+    expect(out.object).toBe("response");
+    expect(out.output).toEqual([{
+      type: "message",
+      role: "assistant",
+      content: [{ type: "output_text", text: "Hey there.", annotations: [] }],
+    }]);
+  });
+
+  it("leaves an OpenAI client's body untouched", () => {
+    const out = translateNonStreamingResponse(KIRO_DECODED_BODY, FORMATS.KIRO, FORMATS.OPENAI);
+
+    expect(out).toBe(KIRO_DECODED_BODY);
+  });
+
+  it("does not touch a body that has no choices array", () => {
+    const raw = { some: "proprietary shape" };
+    const out = translateNonStreamingResponse(raw, FORMATS.KIRO, FORMATS.CLAUDE);
+
+    expect(out).toBe(raw);
+  });
+});