From 9708feddc90a1f26586c8e23b985b7ff02e28840 Mon Sep 17 00:00:00 2001
From: Diego Rodrigues de Sa e Souza
 <8016841+diegosouzapw@users.noreply.github.com>
Date: Sat, 20 Jun 2026 20:25:50 -0300
Subject: [PATCH] fix(translator): strip Claude output_config before MiniMax
 (#4448)

Integrated into release/v3.8.32
---
 CHANGELOG.md                                  |  1 +
 open-sse/translator/helpers/claudeHelper.ts   | 18 +++++++
 ...laude-helper-minimax-output-config.test.ts | 53 +++++++++++++++++++
 3 files changed, 72 insertions(+)
 create mode 100644 tests/unit/claude-helper-minimax-output-config.test.ts

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 24180fbd2f..8ed8ef1584 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -19,6 +19,7 @@ _In development — bullets added per PR; finalized at release._
 - **fix(usage):** reuse Gemini CLI project ID for quota checks (avoid re-discovery). (thanks @Delcado19)
 - **fix(dashboard):** surface manual config CTA when Claude CLI detection fails (remote deployments). (thanks @anuragg-saxenaa)
 - **fix(executors):** granular reasoning_effort handling for Claude models on GitHub Copilot. (thanks @baslr)
+- **fix(translator):** strip Claude output_config before MiniMax (rejected upstream). (thanks @hiepau1231)
 - **fix(combo): round-robin members fail over faster under concurrency saturation via a configurable queue depth** — when a round-robin combo member was saturated, requests sat in the per-model semaphore's **unbounded** queue and only failed over to the next member after the full `queueTimeoutMs` (default 30s) elapsed — so a burst of agentic requests deep-queued one hot member instead of spilling to healthy ones. The per-model semaphore now accepts a bounded queue depth and emits `SEMAPHORE_QUEUE_FULL` once it is full (the round-robin loop already cascades on that code), so a configured low depth fails over immediately. A new `queueDepth` combo-config knob (global default / provider override / per-combo, default **20** for backward compatibility; **0** = never queue → fail over now) is exposed in Settings → Combo Defaults. ([#3872](https://github.com/diegosouzapw/OmniRoute/issues/3872) — thanks @KooshaPari)
 - **fix(pricing): align Claude Code (`cc`) pricing with current Anthropic per-MTok rates** — the `cc` provider block in the default pricing table had stale numbers across every Claude 4.x family entry — most visibly, `claude-opus-4-5-20251101` was billed at the deprecated Opus 4.1 rate (`input $15` / `output $75`), and `claude-haiku-4-5-20251001` was at half the current Haiku 4.5 rate. The `cached` (cache hit) and `cache_creation` (5-minute cache write) multipliers were also off across Opus 4.6/4.7/4.8, Sonnet 4.5/4.6, Haiku 4.5, and Fable 5. All eight entries now match the rates Anthropic publishes (input, 5m cache write at 1.25x input, cache hit at 0.1x input, output; reasoning billed at the output rate), so cost accounting on the dashboard and per-request usage events stop under- or over-reporting Claude Code spend. (thanks @chulanpro5)
 
diff --git a/open-sse/translator/helpers/claudeHelper.ts b/open-sse/translator/helpers/claudeHelper.ts
index 088c354b67..4c5a60e08b 100644
--- a/open-sse/translator/helpers/claudeHelper.ts
+++ b/open-sse/translator/helpers/claudeHelper.ts
@@ -2,6 +2,17 @@
 import { DEFAULT_THINKING_CLAUDE_SIGNATURE } from "../../config/defaultThinkingSignature.ts";
 import { lookupReasoning, recordReplay } from "../../services/reasoningCache.ts";
 
+// MiniMax exposes a Claude-compatible endpoint but rejects Anthropic's extended
+// `output_config` parameter (used to steer reasoning effort and structured output)
+// with a generic 400 "invalid params" response. Strip the entire field before
+// dispatching Claude-shape requests to these providers. Anthropic Claude and
+// other Claude-compatible upstreams that do accept it are unaffected.
+// Ported from upstream decolua/9router#820 by @hiepau1231.
+const CLAUDE_FORMAT_PROVIDERS_WITHOUT_OUTPUT_CONFIG = new Set<string>([
+  "minimax",
+  "minimax-cn",
+]);
+
 // Placeholder thinking text used as last-resort fallback when:
 //   - Target upstream is a non-Anthropic Claude-shape provider
 //     (kimi-coding, glmt, zai, …) that rejects redacted_thinking blobs
@@ -211,6 +222,13 @@ export function prepareClaudeRequest(
   provider: string | null = null,
   preserveCacheControl = false
 ): ClaudeRequestBody {
+  // 0. Strip Anthropic `output_config` for providers that reject it on their
+  // Claude-compatible endpoints (MiniMax). Must run before any downstream
+  // processing so the field never reaches translateRequest/the executor.
+  if (provider && CLAUDE_FORMAT_PROVIDERS_WITHOUT_OUTPUT_CONFIG.has(provider)) {
+    delete body.output_config;
+  }
+
   // 1. System: remove all cache_control, add only to last block with ttl 1h
   // In passthrough mode, preserve existing cache_control markers
   const supportsPromptCaching =
diff --git a/tests/unit/claude-helper-minimax-output-config.test.ts b/tests/unit/claude-helper-minimax-output-config.test.ts
new file mode 100644
index 0000000000..cb0d701aaa
--- /dev/null
+++ b/tests/unit/claude-helper-minimax-output-config.test.ts
@@ -0,0 +1,53 @@
+// Port of upstream decolua/9router#820 by @hiepau1231.
+// MiniMax exposes a Claude-compatible endpoint but rejects Anthropic's
+// extended `output_config` parameter (used to steer reasoning effort and
+// structured output) with a generic 400 "invalid params" response.
+// `prepareClaudeRequest()` must strip the entire `output_config` for
+// MiniMax providers, while preserving it verbatim for Anthropic Claude
+// and other Claude-compatible upstreams that already accept it.
+import { describe, test } from "node:test";
+import assert from "node:assert/strict";
+import { prepareClaudeRequest } from "../../open-sse/translator/helpers/claudeHelper.ts";
+
+describe("prepareClaudeRequest output_config stripping for MiniMax", () => {
+  const buildBody = () => ({
+    model: "MiniMax-M2.7",
+    system: [{ type: "text", text: "You are helpful." }],
+    messages: [{ role: "user", content: [{ type: "text", text: "continue" }] }],
+    max_tokens: 1024,
+    output_config: {
+      effort: "medium",
+      format: {
+        type: "json_schema",
+        schema: {
+          type: "object",
+          properties: { title: { type: "string" } },
+          required: ["title"],
+          additionalProperties: false,
+        },
+      },
+    },
+  });
+
+  test("strips output_config (effort + format) for minimax", () => {
+    const body = buildBody();
+    const result = prepareClaudeRequest(body as any, "minimax");
+    assert.equal(result.output_config, undefined);
+    // Sanity: rest of the request must still be intact.
+    assert.equal(result.messages?.[0]?.content?.[0]?.text, "continue");
+    assert.equal(result.max_tokens, 1024);
+  });
+
+  test("strips output_config (effort + format) for minimax-cn", () => {
+    const body = buildBody();
+    const result = prepareClaudeRequest(body as any, "minimax-cn");
+    assert.equal(result.output_config, undefined);
+  });
+
+  test("preserves output_config for Anthropic Claude", () => {
+    const body = buildBody();
+    const original = JSON.parse(JSON.stringify(body.output_config));
+    const result = prepareClaudeRequest(body as any, "claude");
+    assert.deepEqual(result.output_config, original);
+  });
+});