feat(opencode): send muse-spark max effort first, fallback once to xhigh

changelog.d/features/12687-opencode-muse-spark-max-effort-fallback.md

@@ -0,0 +1 @@
+- **feat(opencode):** send Muse Spark `max` reasoning effort verbatim first (instead of pre-emptively clamping to `xhigh`), retrying once with `xhigh` only when upstream rejects it with the specific unsupported-effort 400 — covers both the fast path and the account-rotation loop ([#12687](https://github.com/diegosouzapw/OmniRoute/pull/12687)) — thanks @Chewji9875

open-sse/config/providerModels.ts

@@ -233,6 +233,20 @@ export function getModelTargetFormat(aliasOrId: string, modelId: string): string
   // and OpenAI-shaped Mistral/Open-MaaS requests. Resource names retain enough publisher data to
   // route future dynamically-synced models without adding another pinned prefix here.
   if (alias === "vertex" || alias === "vp") return getVertexModelTargetFormat(bareModelId);
+  // #10867: Muse Spark is served by OpenCode / Zen / Go / Freebuff only on the
+  // OpenAI Responses API (/responses), not /chat/completions.
+  if (
+    (alias === "opencode" ||
+      alias === "opencode-go" ||
+      alias === "opencode-zen" ||
+      alias === "oc" ||
+      alias === "ocg" ||
+      alias === "ocz" ||
+      alias === "freebuff") &&
+    /(?:^|\/)muse-spark/i.test(bareModelId)
+  ) {
+    return "openai-responses";
+  }
   // Model-level targetFormat is provider-scoped: a catalog entry declares how THIS
   // provider's endpoint serves the model — do NOT import another provider's tag.
   // #9994 scoped this for providers WITH a catalog; #10072 extends it to catalogless

open-sse/config/providers/registry/command-code/index.ts

@@ -184,5 +184,23 @@ export const command_codeProvider: RegistryEntry = {
       contextLength: 1000000,
       maxOutputTokens: 32768,
     },
+    {
+      id: "meta/muse-spark-1.3-contributor",
+      name: "Muse Spark 1.3 Contributor (CC)",
+      supportsReasoning: true,
+      supportedThinkingEfforts: COMMAND_CODE_REASONING_EFFORTS,
+      supportsVision: true,
+      contextLength: 1048576,
+      maxOutputTokens: 131072,
+    },
+    {
+      id: "meta/muse-spark-1.2-contributor",
+      name: "Muse Spark 1.2 Contributor (CC)",
+      supportsReasoning: true,
+      supportedThinkingEfforts: COMMAND_CODE_REASONING_EFFORTS,
+      supportsVision: true,
+      contextLength: 1048576,
+      maxOutputTokens: 131072,
+    },
   ],
 };

open-sse/config/providers/registry/freebuff/index.ts

@@ -60,6 +60,12 @@ export const freebuffProvider: RegistryEntry = {
       supportsReasoning: true,
       contextLength: 131_072,
     },
+    {
+      id: "meta/muse-spark-1.3-contributor",
+      name: "Meta Muse Spark 1.3 Contributor",
+      supportsReasoning: true,
+      contextLength: 131_072,
+    },
     {
       id: "meta/muse-spark-1.2-contributor",
       name: "Meta Muse Spark 1.2 Contributor",

open-sse/config/providers/registry/opencode/go/index.ts

@@ -247,7 +247,9 @@ export const opencode_goProvider: RegistryEntry = {
     },
     // #12674: Muse Spark 1.3 Contributor — base + effort-tier aliases from the
     // OpenCode Go registry (`opencode models opencode-go --refresh --verbose`;
-    // exact suffix set: minimal/low/medium/high/xhigh, no max — same as 1.2).
+    // exact suffix set: minimal/low/medium/high/xhigh — same as 1.2). #12687 adds
+    // a `-max` variant: the executor sends `max` verbatim and retries once with
+    // `xhigh` only if upstream rejects it (dispatchWithMuseSparkMaxFallback).
     // Upstream serves Muse Spark only on the Responses API; without
     // targetFormat:"openai-responses" these fall through to /chat/completions
     // and the upstream returns 500 (same class as #12196).
@@ -317,6 +319,17 @@ export const opencode_goProvider: RegistryEntry = {
       supportsVideo: true,
       targetFormat: "openai-responses",
     },
+    {
+      id: "muse-spark-1.3-contributor-max",
+      name: "Muse Spark 1.3 Contributor (max effort)",
+      contextLength: 1048576,
+      maxOutputTokens: 131072,
+      supportsReasoning: true,
+      supportsVision: true,
+      supportsAudio: true,
+      supportsVideo: true,
+      targetFormat: "openai-responses",
+    },
     // #8353: Grok 4.5 + effort tiers from the OpenCode Go registry.
     {
       id: "grok-4.5",

open-sse/executors/base/reasoningEffort.ts

@@ -62,6 +62,7 @@ export const MAX_TIER_REASONING_MODEL_PATTERN =
 
 export const O1_O3_REASONING_MODELS_PATTERN = /(?:^|\/|\b)(?:o1-mini|o1|o3-mini|o3-pro|o3)(?:$|-)/i;
 export const O1_PREVIEW_PATTERN = /(?:^|\/|\b)o1-preview(?:$|-)/i;
+export const MUSE_SPARK_12_PATTERN = /(?:^|\/|\b)muse-spark-1\.2(?:$|-|\b)/i;
 export const MUSE_SPARK_PATTERN = /(?:^|\/|\b)muse-spark/i;
 export const MINIMAX_REASONING_PATTERN = /(?:^|\/|\b)minimax(?:-m3|-m2)/i;
 export const GROK_45_PATTERN = /(?:^|\/|\b)grok-4\.5/i;
@@ -70,19 +71,11 @@ export const GLM_53_FAMILY_PATTERN = /(?:^|\/|\b)glm-5\.3(?:$|-)/i;
 export const GLM_52_FAMILY_PATTERN = /(?:^|\/|\b)glm-5\.2(?:$|-)/i;
 
 export function isCommandCodeProvider(provider: string): boolean {
-  return (
-    provider === "command-code" ||
-    provider === "cmd" ||
-    provider === "command_code"
-  );
+  return provider === "command-code" || provider === "cmd" || provider === "command_code";
 }
 
 export function isOllamaCloudProvider(provider: string): boolean {
-  return (
-    provider === "ollama-cloud" ||
-    provider === "ollamacloud" ||
-    provider === "ollama_cloud"
-  );
+  return provider === "ollama-cloud" || provider === "ollamacloud" || provider === "ollama_cloud";
 }
 
 export function isOpencodeGoProvider(provider: string): boolean {
@@ -206,12 +199,7 @@ export function supportsMaxEffortForProvider(provider: string, model: string): b
     MAX_TIER_REASONING_MODEL_PATTERN.test(resolvedModelId) ||
     MAX_TIER_REASONING_MODEL_PATTERN.test(model);
   return (
-    isClaude ||
-    isOpencodeGo ||
-    isOllamaCloud ||
-    isMoonshotK3 ||
-    isCommandCode ||
-    isMaxTierModel
+    isClaude || isOpencodeGo || isOllamaCloud || isMoonshotK3 || isCommandCode || isMaxTierModel
   );
 }
 
@@ -390,14 +378,16 @@ export function sanitizeReasoningEffortForProvider(
     return body;
   }
 
-  // ── Muse Spark models (muse-spark-1.2, etc.) ─────────────────────────────
-  // Accepts minimal|low|medium|high|xhigh. Rejects none (400) and max.
-  // max/ultra → xhigh; none → minimal.
-  if (MUSE_SPARK_PATTERN.test(modelStr)) {
+  // ── Muse Spark models (muse-spark-1.2 vs 1.3+ / future versions) ───────
+  // 1.2: Accepts minimal|low|medium|high|xhigh (no max).
+  //      max/ultra → xhigh; none → minimal.
+  // 1.3+ / future versions: Accepts minimal|low|medium|high|xhigh|max.
+  //      ultra → max; none → minimal.
+  if (MUSE_SPARK_12_PATTERN.test(modelStr)) {
     if (effortStr === "max" || effortStr === "ultra") {
       log?.info?.(
         "REASONING_SANITIZE",
-        `${provider}/${modelStr}: clamped reasoning_effort ${effortStr} → xhigh (Muse Spark ceiling)`
+        `${provider}/${modelStr}: clamped reasoning_effort ${effortStr} → xhigh (Muse Spark 1.2 ceiling)`
       );
       return writeEffortValue(b, "xhigh", c);
     }
@@ -411,6 +401,24 @@ export function sanitizeReasoningEffortForProvider(
     return body;
   }
 
+  if (MUSE_SPARK_PATTERN.test(modelStr)) {
+    if (effortStr === "ultra") {
+      log?.info?.(
+        "REASONING_SANITIZE",
+        `${provider}/${modelStr}: clamped reasoning_effort ultra → max (Muse Spark ceiling)`
+      );
+      return writeEffortValue(b, "max", c);
+    }
+    if (effortStr === "none") {
+      log?.info?.(
+        "REASONING_SANITIZE",
+        `${provider}/${modelStr}: clamped reasoning_effort none → minimal (Muse Spark floor)`
+      );
+      return writeEffortValue(b, "minimal", c);
+    }
+    return body;
+  }
+
   // ── OpenAI o1 / o3-mini models ───────────────────────────────────────────
   // Accepts only low|medium|high. Clamp xhigh/max/ultra → high.
   if (O1_O3_REASONING_MODELS_PATTERN.test(modelStr)) {
@@ -549,11 +557,10 @@ export function sanitizeReasoningEffortForProvider(
     ? modelStr.slice(provider.length + 1)
     : modelStr;
   const declaredEfforts = getProviderModels(provider).find(
-    (entry) => entry.id === providerModelIdForClamp || entry.aliases?.includes(providerModelIdForClamp)
+    (entry) =>
+      entry.id === providerModelIdForClamp || entry.aliases?.includes(providerModelIdForClamp)
   )?.supportedThinkingEfforts;
-  const declaredRanked = (
-    Array.isArray(declaredEfforts) ? declaredEfforts : []
-  )
+  const declaredRanked = (Array.isArray(declaredEfforts) ? declaredEfforts : [])
     .map((tier) => ({ tier, rank: REASONING_EFFORT_ORDER.indexOf(tier) }))
     .filter((x) => x.rank >= 0)
     .sort((a, b) => a.rank - b.rank);

open-sse/executors/freebuff.ts

@@ -12,6 +12,7 @@ const MODEL_TO_AGENT: Record<string, string> = {
   "z-ai/glm-5.2": "base2-free-glm",
   "crof/kimi-k3-eco": "base2-free-kimi-k3-eco",
   "anthropic/claude-fable-5": "base2-free-fable",
+  "meta/muse-spark-1.3-contributor": "base2-free-muse-spark",
   "meta/muse-spark-1.2-contributor": "base2-free-muse-spark",
 };
 

open-sse/executors/opencode.ts

@@ -65,6 +65,13 @@ import {
   isResponsesTerminalLine,
   normalizeMuseSparkFinishReason,
 } from "./opencodeMuseSpark.ts";
+import {
+  applyMuseSparkResponsesEffort,
+  dispatchWithMuseSparkMaxFallback,
+  parseDynamicMuseSparkEffort,
+  type HttpExecuteResult,
+} from "./opencodeMuseSparkEffort.ts";
+export { isUnsupportedReasoningEffortRejection } from "./opencodeMuseSparkEffort.ts";
 import { currentRequestContext, runInRequestContext } from "./opencodeRequestContext.ts";
 import {
   handleLoopFreeTierRefusal,
@@ -138,8 +145,11 @@ const EFFORT_LEVELS = ["none", "low", "high", "max"] as const;
  *   grok-4.5 low/medium/high; hy3 none/low/high; kimi-k3 max;
  *   qwen3.6-plus / qwen3.7-max / qwen3.7-plus high/max;
  *   muse-spark-1.2-contributor minimal/low/medium/high/xhigh (no max)
- * - #12674 Muse Spark 1.3 Contributor: minimal/low/medium/high/xhigh (no max),
- *   verified via `opencode models opencode-go --refresh --verbose`
+ * - #12674 Muse Spark 1.3 Contributor: minimal/low/medium/high/xhigh verified
+ *   via `opencode models opencode-go --refresh --verbose`; #12687 adds `max`
+ *   as an explicit alias suffix — the wire tier is still sent verbatim and
+ *   falls back to xhigh only on the upstream's unsupported-effort 400 (see
+ *   dispatchWithMuseSparkMaxFallback).
  */
 const EFFORT_TIERS: Record<string, readonly string[]> = {
   "deepseek-v4-pro": EFFORT_LEVELS,
@@ -152,8 +162,8 @@ const EFFORT_TIERS: Record<string, readonly string[]> = {
   "qwen3.6-plus": ["high", "max"],
   "qwen3.7-max": ["high", "max"],
   "qwen3.7-plus": ["high", "max"],
+  "muse-spark-1.3-contributor": ["minimal", "low", "medium", "high", "xhigh", "max"],
   "muse-spark-1.2-contributor": ["minimal", "low", "medium", "high", "xhigh"],
-  "muse-spark-1.3-contributor": ["minimal", "low", "medium", "high", "xhigh"],
 };
 
 /**
@@ -171,7 +181,7 @@ export function parseEffortLevel(model: string): { baseModel: string; effort: st
       }
     }
   }
-  return null;
+  return parseDynamicMuseSparkEffort(m);
 }
 
 /**
@@ -543,9 +553,14 @@ export class OpencodeExecutor extends BaseExecutor {
         // execute() in runWithProxyContext(proxyInfo.proxy, ...) before we run.
         // Only pin direct egress when no such context exists; otherwise let the
         // ambient proxy stand instead of clobbering it with the direct sentinel.
-        const dispatch = () => super.execute(input);
+        const dispatch = (effInput: ExecuteInput) =>
+          super.execute(effInput) as Promise<HttpExecuteResult>;
         const single = (await guardStall(
-          await (hasAmbientProxyContext() ? dispatch() : runWithDirectFetchContext(dispatch))
+          await (hasAmbientProxyContext()
+            ? dispatchWithMuseSparkMaxFallback(input, dispatch, parseEffortLevel)
+            : runWithDirectFetchContext(() =>
+                dispatchWithMuseSparkMaxFallback(input, dispatch, parseEffortLevel)
+              ))
         )) as HttpExecuteResult;
         const retryAfterRefusal = await retryFreeTierRefusalWithObservedTools(
           this.freeTierRetryCtx(input),
@@ -597,10 +612,6 @@ export class OpencodeExecutor extends BaseExecutor {
       // This loop only ever dispatches through super.execute() (the HTTP request
       // path), which always resolves the object-shaped arm of ExecutorExecuteResult
       // — the bare-Response arm belongs to web/scraping executors only (base.ts:290).
-      type HttpExecuteResult = Extract<
-        Awaited<ReturnType<BaseExecutor["execute"]>>,
-        { response: Response }
-      >;
       let lastResult: HttpExecuteResult | null = null;
       let lastSharedEgressError: unknown = null;
       const sharedEgressGuardEnabled = isNetworkRotationSharedEgressGuardEnabled();
@@ -749,10 +760,12 @@ export class OpencodeExecutor extends BaseExecutor {
         let result: HttpExecuteResult;
         try {
           // super.execute() dispatches the HTTP path (never the web/scraping arm).
+          const loopDispatch = (effInput: ExecuteInput) =>
+            runWithProxyContext(account.proxy, () =>
+              super.execute({ ...effInput, skipUpstreamRetry: true })
+            ) as Promise<HttpExecuteResult>;
           result = (await guardStall(
-            await runWithProxyContext(account.proxy, () =>
-              super.execute({ ...input, skipUpstreamRetry: true })
-            )
+            await dispatchWithMuseSparkMaxFallback(input, loopDispatch, parseEffortLevel)
           )) as HttpExecuteResult;
         } catch (err) {
           const reason = err instanceof Error ? err.message : String(err);
@@ -1272,13 +1285,16 @@ export class OpencodeExecutor extends BaseExecutor {
       if (parsed) {
         const deepseekFamily =
           parsed.baseModel === "deepseek-v4-pro" || parsed.baseModel === "deepseek-v4-flash";
+        const museSparkFamily = parsed.baseModel.startsWith("muse-spark");
         if (deepseekFamily) {
           // DeepSeek via opencode-go proxies the native DeepSeek contract, which
           // accepts a flat reasoning_effort field (#4647).
           mb.model = parsed.baseModel;
           if (mb.reasoning_effort === undefined) {
             mb.reasoning_effort = parsed.effort;
           }
+        } else if (museSparkFamily) {
+          applyMuseSparkResponsesEffort(mb, parsed);
         }
         // #10788: every other family's ONLY native effort mechanism is the
         // -<tier> suffix in the model id itself (the ids `opencode models