fix(sse): recognize xiaomi-tokenplan mimo as a thinking-mode model (#7098)

* fix(sse): recognize xiaomi-tokenplan mimo as a thinking-mode model (port from 9router#1321)

The reasoning_content injector already handles DeepSeek/Kimi/K2/MiniMax thinking-mode upstreams, echoing a placeholder reasoning_content on assistant turns that lack one. Its THINKING_MODEL_PATTERNS list omitted the xiaomi-tokenplan mimo family, so requests through xiaomi-tokenplan/mimo-v2.5-pro still hit upstream's 400 'reasoning_content in the thinking mode must be passed back to the API', making the model unusable in multi-turn conversations (e.g. Codex CLI). Add a /\bmimo\b/i pattern so mimo models get the same treatment.

Reported-by: z.wl (@xxue-z) (https://github.com/decolua/9router/issues/1321)

* docs(changelog): add fragment for #7098 mimo thinking-model fix

changelog.d/fixes/7098-mimo-thinking-model-reasoning-content.md

@@ -0,0 +1 @@
+- **fix(sse):** xiaomi-tokenplan `mimo` models (e.g. `mimo-v2.5-pro`) are now recognized as thinking-mode upstreams that require `reasoning_content` echoed back on every assistant turn, fixing a persistent `400 reasoning_content must be passed back` error on multi-turn conversations ([#7098](https://github.com/diegosouzapw/OmniRoute/pull/7098)) — thanks @xxue-z

open-sse/utils/reasoningContentInjector.ts

@@ -1,5 +1,6 @@
 /**
- * Thinking-mode upstreams (DeepSeek V4 Flash, Kimi, MiniMax, ...) require
+ * Thinking-mode upstreams (DeepSeek V4 Flash, Kimi, MiniMax, xiaomi-tokenplan
+ * mimo, ...) require
  * `reasoning_content` to be echoed back on every assistant message in the
  * conversation history. Standard OpenAI clients do not preserve that field
  * across turns, so we inject a non-empty placeholder before forwarding.
@@ -26,6 +27,7 @@ const THINKING_MODEL_PATTERNS: RegExp[] = [
   /\bkimi\b/i,
   /\bk2\b/i, // moonshot kimi k2 family alias
   /\bminimax\b/i,
+  /\bmimo\b/i, // xiaomi-tokenplan mimo family (e.g. xiaomi-tokenplan/mimo-v2.5-pro)
 ];
 
 export function isThinkingMessageModel(model: string | undefined | null): boolean {

tests/unit/reasoningContentInjector.test.ts

@@ -0,0 +1,44 @@
+import { describe, it } from "node:test";
+import assert from "node:assert/strict";
+
+import {
+  isThinkingMessageModel,
+  injectReasoningContentForThinkingModel,
+} from "../../open-sse/utils/reasoningContentInjector.ts";
+
+describe("reasoningContentInjector — xiaomi-tokenplan mimo family (9router#1321)", () => {
+  it("recognizes xiaomi-tokenplan/mimo-v2.5-pro as a thinking-mode model", () => {
+    assert.equal(isThinkingMessageModel("xiaomi-tokenplan/mimo-v2.5-pro"), true);
+  });
+
+  it("recognizes bare mimo model ids as thinking-mode models", () => {
+    assert.equal(isThinkingMessageModel("mimo-v2.5-pro"), true);
+  });
+
+  it("still recognizes the existing thinking-mode families (deepseek/kimi/k2/minimax)", () => {
+    assert.equal(isThinkingMessageModel("deepseek-v4-flash"), true);
+    assert.equal(isThinkingMessageModel("kimi-k2"), true);
+    assert.equal(isThinkingMessageModel("minimax-m2"), true);
+  });
+
+  it("does not flag unrelated model ids", () => {
+    assert.equal(isThinkingMessageModel("gpt-4o"), false);
+  });
+
+  it("injects a reasoning_content placeholder for assistant messages when routed to mimo", () => {
+    const body = {
+      model: "xiaomi-tokenplan/mimo-v2.5-pro",
+      messages: [
+        { role: "user", content: "hi" },
+        { role: "assistant", content: "hello" },
+      ],
+    };
+
+    // Simulate the executor gate: only inject when the model is a thinking model.
+    assert.equal(isThinkingMessageModel(body.model), true);
+
+    const result = injectReasoningContentForThinkingModel(body) as typeof body;
+    const assistantMsg = result.messages[1] as Record<string, unknown>;
+    assert.equal(assistantMsg.reasoning_content, " ");
+  });
+});