remote-codex 0.11.30 → 0.11.31

package/apps/supervisor-api/dist/index.js

@@ -291,19 +291,35 @@ function parsePricingConfig(raw) {
     if (!isPositiveNumber(entry.inputUsdPerMillion) || !isPositiveNumber(entry.cachedInputUsdPerMillion) || !isPositiveNumber(entry.outputUsdPerMillion) || typeof entry.supportsFastMode !== "boolean") {
       throw new Error(`Pricing config model "${modelKey2}" has invalid fields.`);
     }
+    if (entry.cacheWriteInputUsdPerMillion !== void 0 && !isPositiveNumber(entry.cacheWriteInputUsdPerMillion)) {
+      throw new Error(`Pricing config model "${modelKey2}" cacheWriteInputUsdPerMillion must be a non-negative number.`);
+    }
     if (entry.fastMultiplier !== void 0 && !isPositiveNumber(entry.fastMultiplier)) {
       throw new Error(`Pricing config model "${modelKey2}" fastMultiplier must be a non-negative number.`);
     }
     if (entry.contextWindowTokens !== void 0 && !isPositiveNumber(entry.contextWindowTokens)) {
       throw new Error(`Pricing config model "${modelKey2}" contextWindowTokens must be a non-negative number.`);
     }
+    for (const field of [
+      "longContextThresholdTokens",
+      "longContextInputMultiplier",
+      "longContextOutputMultiplier"
+    ]) {
+      if (entry[field] !== void 0 && !isPositiveNumber(entry[field])) {
+        throw new Error(`Pricing config model "${modelKey2}" ${field} must be a non-negative number.`);
+      }
+    }
     models[modelKey2] = {
       inputUsdPerMillion: entry.inputUsdPerMillion,
       cachedInputUsdPerMillion: entry.cachedInputUsdPerMillion,
+      ...entry.cacheWriteInputUsdPerMillion !== void 0 ? { cacheWriteInputUsdPerMillion: entry.cacheWriteInputUsdPerMillion } : {},
       outputUsdPerMillion: entry.outputUsdPerMillion,
       supportsFastMode: entry.supportsFastMode,
       ...entry.fastMultiplier !== void 0 ? { fastMultiplier: entry.fastMultiplier } : {},
-      ...entry.contextWindowTokens !== void 0 ? { contextWindowTokens: entry.contextWindowTokens } : {}
+      ...entry.contextWindowTokens !== void 0 ? { contextWindowTokens: entry.contextWindowTokens } : {},
+      ...entry.longContextThresholdTokens !== void 0 ? { longContextThresholdTokens: entry.longContextThresholdTokens } : {},
+      ...entry.longContextInputMultiplier !== void 0 ? { longContextInputMultiplier: entry.longContextInputMultiplier } : {},
+      ...entry.longContextOutputMultiplier !== void 0 ? { longContextOutputMultiplier: entry.longContextOutputMultiplier } : {}
     };
   }
   return {
@@ -396,23 +412,32 @@ function estimateTurnPrice(usage, snapshot) {
     return null;
   }
   const nonCachedInputTokens = Math.max(
-    usage.total.inputTokens - usage.total.cachedInputTokens,
+    usage.total.inputTokens - usage.total.cachedInputTokens - (usage.total.cacheWriteInputTokens ?? 0),
     0
   );
   const cachedInputTokens = Math.max(usage.total.cachedInputTokens, 0);
+  const cacheWriteInputTokens = Math.max(
+    usage.total.cacheWriteInputTokens ?? 0,
+    0
+  );
   const outputTokens = Math.max(usage.total.outputTokens, 0);
-  const multiplier = tierKey === "fast" && modelPricing.fastMultiplier !== void 0 ? modelPricing.fastMultiplier : tier.multiplier;
-  const inputUsd = nonCachedInputTokens * modelPricing.inputUsdPerMillion * multiplier / TOKEN_PRICE_DENOMINATOR;
-  const cachedInputUsd = cachedInputTokens * modelPricing.cachedInputUsdPerMillion * multiplier / TOKEN_PRICE_DENOMINATOR;
-  const outputUsd = outputTokens * modelPricing.outputUsdPerMillion * multiplier / TOKEN_PRICE_DENOMINATOR;
+  const tierMultiplier = tierKey === "fast" && modelPricing.fastMultiplier !== void 0 ? modelPricing.fastMultiplier : tier.multiplier;
+  const usesLongContextPricing = typeof modelPricing.longContextThresholdTokens === "number" && usage.total.inputTokens > modelPricing.longContextThresholdTokens;
+  const inputMultiplier = tierMultiplier * (usesLongContextPricing ? modelPricing.longContextInputMultiplier ?? 1 : 1);
+  const outputMultiplier = tierMultiplier * (usesLongContextPricing ? modelPricing.longContextOutputMultiplier ?? 1 : 1);
+  const inputUsd = nonCachedInputTokens * modelPricing.inputUsdPerMillion * inputMultiplier / TOKEN_PRICE_DENOMINATOR;
+  const cachedInputUsd = cachedInputTokens * modelPricing.cachedInputUsdPerMillion * inputMultiplier / TOKEN_PRICE_DENOMINATOR;
+  const cacheWriteInputUsd = cacheWriteInputTokens * (modelPricing.cacheWriteInputUsdPerMillion ?? modelPricing.inputUsdPerMillion) * inputMultiplier / TOKEN_PRICE_DENOMINATOR;
+  const outputUsd = outputTokens * modelPricing.outputUsdPerMillion * outputMultiplier / TOKEN_PRICE_DENOMINATOR;
   return {
     pricingModelKey,
     pricingTierKey: tierKey,
     currency: pricingConfig.currency,
     inputUsd,
     cachedInputUsd,
+    cacheWriteInputUsd,
     outputUsd,
-    totalUsd: inputUsd + cachedInputUsd + outputUsd
+    totalUsd: inputUsd + cachedInputUsd + cacheWriteInputUsd + outputUsd
   };
 }
 
@@ -11970,6 +11995,7 @@ function addClaudeUsage(left, right) {
     totalTokens: left.totalTokens + right.totalTokens,
     inputTokens: left.inputTokens + right.inputTokens,
     cachedInputTokens: left.cachedInputTokens + right.cachedInputTokens,
+    cacheWriteInputTokens: left.cacheWriteInputTokens + right.cacheWriteInputTokens,
     outputTokens: left.outputTokens + right.outputTokens,
     reasoningOutputTokens: left.reasoningOutputTokens + right.reasoningOutputTokens
   };
@@ -11995,6 +12021,7 @@ function normalizeClaudeUsage(value) {
     totalTokens: totalTokens2,
     inputTokens,
     cachedInputTokens: cacheReadInputTokens,
+    cacheWriteInputTokens: cacheCreationInputTokens,
     outputTokens,
     reasoningOutputTokens: 0
   };
@@ -14278,6 +14305,9 @@ function openCodeUsageFromTokens(tokens) {
   const cachedInputTokens = nonNegativeNumberValue(
     tokens.cachedInputTokens ?? tokens.cached_input_tokens ?? cache?.read
   ) ?? 0;
+  const cacheWriteInputTokens = nonNegativeNumberValue(
+    tokens.cacheWriteInputTokens ?? tokens.cache_write_input_tokens ?? tokens.cacheWriteTokens ?? tokens.cache_write_tokens ?? cache?.write
+  ) ?? 0;
   const totalTokens2 = nonNegativeNumberValue(tokens.total ?? tokens.totalTokens ?? tokens.total_tokens) ?? inputTokens + outputTokens;
   if (totalTokens2 <= 0) {
     return null;
@@ -14287,6 +14317,7 @@ function openCodeUsageFromTokens(tokens) {
       totalTokens: totalTokens2,
       inputTokens,
       cachedInputTokens,
+      ...cacheWriteInputTokens > 0 ? { cacheWriteInputTokens } : {},
       outputTokens,
       reasoningOutputTokens
     },
@@ -14344,6 +14375,9 @@ function turnTokenUsage(messages, model) {
     totalTokens: sum.totalTokens + usage.totalTokens,
     inputTokens: sum.inputTokens + usage.inputTokens,
     cachedInputTokens: sum.cachedInputTokens + usage.cachedInputTokens,
+    ...(sum.cacheWriteInputTokens ?? 0) + (usage.cacheWriteInputTokens ?? 0) > 0 ? {
+      cacheWriteInputTokens: (sum.cacheWriteInputTokens ?? 0) + (usage.cacheWriteInputTokens ?? 0)
+    } : {},
     outputTokens: sum.outputTokens + usage.outputTokens,
     reasoningOutputTokens: sum.reasoningOutputTokens + usage.reasoningOutputTokens
   }), firstRecord);
@@ -17311,11 +17345,18 @@ function shouldResetThreadContextUsageForTurnStart(current) {
 }
 function buildTurnTokenBreakdown(payload) {
   const usage = isRecord12(payload) ? payload : null;
+  const inputDetails = isRecord12(
+    usage?.inputTokensDetails ?? usage?.input_tokens_details
+  ) ? usage?.inputTokensDetails ?? usage?.input_tokens_details : null;
+  const cache = isRecord12(usage?.cache) ? usage.cache : null;
   const totalTokens2 = numberOrNull2(usage?.totalTokens ?? usage?.total_tokens);
   const inputTokens = numberOrNull2(usage?.inputTokens ?? usage?.input_tokens);
   const cachedInputTokens = numberOrNull2(
-    usage?.cachedInputTokens ?? usage?.cached_input_tokens
+    usage?.cachedInputTokens ?? usage?.cached_input_tokens ?? inputDetails?.cachedTokens ?? inputDetails?.cached_tokens ?? cache?.read
   );
+  const cacheWriteInputTokens = numberOrNull2(
+    usage?.cacheWriteInputTokens ?? usage?.cache_write_input_tokens ?? usage?.cacheWriteTokens ?? usage?.cache_write_tokens ?? usage?.cacheCreationInputTokens ?? usage?.cache_creation_input_tokens ?? inputDetails?.cacheWriteTokens ?? inputDetails?.cache_write_tokens ?? cache?.write
+  ) ?? 0;
   const outputTokens = numberOrNull2(usage?.outputTokens ?? usage?.output_tokens);
   const reasoningOutputTokens = numberOrNull2(
     usage?.reasoningOutputTokens ?? usage?.reasoning_output_tokens
@@ -17327,6 +17368,7 @@ function buildTurnTokenBreakdown(payload) {
     totalTokens: totalTokens2,
     inputTokens,
     cachedInputTokens,
+    cacheWriteInputTokens,
     outputTokens,
     reasoningOutputTokens
   };
@@ -17336,6 +17378,7 @@ function zeroTurnTokenBreakdown() {
     totalTokens: 0,
     inputTokens: 0,
     cachedInputTokens: 0,
+    cacheWriteInputTokens: 0,
     outputTokens: 0,
     reasoningOutputTokens: 0
   };
@@ -17348,6 +17391,10 @@ function subtractTurnTokenBreakdowns(current, previous) {
       current.cachedInputTokens - previous.cachedInputTokens,
       0
     ),
+    cacheWriteInputTokens: Math.max(
+      current.cacheWriteInputTokens - previous.cacheWriteInputTokens,
+      0
+    ),
     outputTokens: Math.max(current.outputTokens - previous.outputTokens, 0),
     reasoningOutputTokens: Math.max(
       current.reasoningOutputTokens - previous.reasoningOutputTokens,
@@ -17406,12 +17453,16 @@ function cumulativeTotalFromStoredThreadTurnTokenUsageState(state) {
     return null;
   }
   if (!state.baselineTotal) {
-    return state.usage.total;
+    return {
+      ...state.usage.total,
+      cacheWriteInputTokens: state.usage.total.cacheWriteInputTokens ?? 0
+    };
   }
   return {
     totalTokens: state.baselineTotal.totalTokens + state.usage.total.totalTokens,
     inputTokens: state.baselineTotal.inputTokens + state.usage.total.inputTokens,
     cachedInputTokens: state.baselineTotal.cachedInputTokens + state.usage.total.cachedInputTokens,
+    cacheWriteInputTokens: state.baselineTotal.cacheWriteInputTokens + (state.usage.total.cacheWriteInputTokens ?? 0),
     outputTokens: state.baselineTotal.outputTokens + state.usage.total.outputTokens,
     reasoningOutputTokens: state.baselineTotal.reasoningOutputTokens + state.usage.total.reasoningOutputTokens
   };
@@ -18763,6 +18814,7 @@ function normalizeReasoningEffort(value) {
     case "high":
     case "xhigh":
     case "max":
+    case "ultra":
       return value;
     default:
       return null;
@@ -18779,6 +18831,7 @@ function normalizeReasoningEffort2(value) {
     case "high":
     case "xhigh":
     case "max":
+    case "ultra":
       return value;
     default:
       return null;
@@ -23519,7 +23572,8 @@ var reasoningEffortValues = [
   "medium",
   "high",
   "xhigh",
-  "max"
+  "max",
+  "ultra"
 ];
 var createThreadSchema = z5.object({
   workspaceId: z5.string().uuid(),