pi-commandcode-provider 0.4.3 → 0.5.1

package/src/models.ts

@@ -2,10 +2,154 @@ import { mkdir, readFile, rename, rm, writeFile } from "node:fs/promises"
 import { dirname } from "node:path"
 
 export const DEFAULT_MODELS_URL = "https://api.commandcode.ai/provider/v1/models"
+export const DEFAULT_MODELS_TIMEOUT_MS = 10_000
 
 const DEFAULT_MAX_OUTPUT_TOKENS = 65_536
 const MODEL_CACHE_VERSION = 1
 
+export type CommandCodeInputType = "text" | "image"
+
+/**
+ * Model input modalities from the command-code@1.15.1 bundled catalog.
+ * Models omitted here remain text-only so newly discovered IDs never claim
+ * image support without upstream evidence.
+ */
+export const MODEL_INPUT_MODALITIES: Readonly<Record<string, readonly CommandCodeInputType[]>> = {
+  "MiniMaxAI/MiniMax-M3": ["text", "image"],
+  "Qwen/Qwen3.6-Plus": ["text", "image"],
+  "Qwen/Qwen3.7-Flash": ["text", "image"],
+  "Qwen/Qwen3.7-Plus": ["text", "image"],
+  "Qwen/Qwen3.8-Max": ["text", "image"],
+  "claude-fable-5": ["text", "image"],
+  "claude-haiku-4-5-20251001": ["text", "image"],
+  "claude-opus-4-7": ["text", "image"],
+  "claude-opus-4-8": ["text", "image"],
+  "claude-opus-5": ["text", "image"],
+  "claude-sonnet-4-6": ["text", "image"],
+  "claude-sonnet-5": ["text", "image"],
+  "google/gemini-3.1-flash-lite": ["text", "image"],
+  "google/gemini-3.5-flash": ["text", "image"],
+  "google/gemini-3.5-flash-lite": ["text", "image"],
+  "google/gemini-3.6-flash": ["text", "image"],
+  "gpt-5.3-codex": ["text", "image"],
+  "gpt-5.4": ["text", "image"],
+  "gpt-5.4-mini": ["text", "image"],
+  "gpt-5.5": ["text", "image"],
+  "gpt-5.6-luna": ["text", "image"],
+  "gpt-5.6-sol": ["text", "image"],
+  "gpt-5.6-terra": ["text", "image"],
+  "meta/muse-spark-1.1": ["text", "image"],
+  "meta/muse-spark-1.2": ["text", "image"],
+  "meta/muse-spark-1.2-contributor": ["text", "image"],
+  "moonshotai/Kimi-K2.5": ["text", "image"],
+  "moonshotai/Kimi-K2.6": ["text", "image"],
+  "moonshotai/Kimi-K2.7-Code": ["text", "image"],
+  "moonshotai/Kimi-K2.7-Code-Highspeed": ["text", "image"],
+  "moonshotai/Kimi-K3": ["text", "image"],
+  "sakana/fugu-ultra": ["text", "image"],
+  "stepfun/Step-3.7-Flash": ["text", "image"],
+  "thinkingmachines/inkling": ["text", "image"],
+  "thinkingmachines/inkling-small": ["text", "image"],
+  "xai/grok-4.5": ["text", "image"],
+  "xiaomi/mimo-v2.5": ["text", "image"],
+}
+
+const TEXT_INPUT_ONLY = ["text"] as const
+
+export function inputModalitiesForModel(modelId: string): readonly CommandCodeInputType[] {
+  return MODEL_INPUT_MODALITIES[modelId] ?? TEXT_INPUT_ONLY
+}
+
+export function modelSupportsImageInput(modelId: string): boolean {
+  return inputModalitiesForModel(modelId).includes("image")
+}
+
+export type PiThinkingLevel = "off" | "minimal" | "low" | "medium" | "high" | "xhigh" | "max"
+
+type CommandCodeReasoningEffort = Exclude<PiThinkingLevel, "off">
+
+/**
+ * Per-model reasoning efforts supported by Command Code's generate endpoint.
+ *
+ * The Provider API does not expose reasoning metadata. This is an exact
+ * snapshot of `reasoningEfforts` from the command-code@1.15.1 model catalog
+ * (`packages/shared/src/model-catalog.ts`, also published in the generated
+ * `dist/bundled/command-code-knowledge/reference/models.md`). Models omitted
+ * here let Command Code choose their reasoning depth, matching the CLI.
+ */
+export const MODEL_EFFORTS: Readonly<Record<string, readonly CommandCodeReasoningEffort[]>> = {
+  "Qwen/Qwen3.8-Max": ["low", "medium", "xhigh"],
+  "claude-fable-5": ["low", "medium", "high", "xhigh", "max"],
+  "claude-opus-4-7": ["low", "medium", "high", "xhigh", "max"],
+  "claude-opus-4-8": ["low", "medium", "high", "xhigh", "max"],
+  "claude-opus-5": ["low", "medium", "high", "xhigh", "max"],
+  "claude-sonnet-4-6": ["low", "medium", "high", "xhigh", "max"],
+  "claude-sonnet-5": ["low", "medium", "high", "xhigh", "max"],
+  "deepseek/deepseek-v4-flash": ["high", "max"],
+  "deepseek/deepseek-v4-pro": ["high", "max"],
+  "gpt-5.3-codex": ["low", "medium", "high", "xhigh"],
+  "gpt-5.4": ["low", "medium", "high", "xhigh"],
+  "gpt-5.4-mini": ["low", "medium", "high"],
+  "gpt-5.5": ["low", "medium", "high", "xhigh"],
+  "gpt-5.6-luna": ["low", "medium", "high", "xhigh", "max"],
+  "gpt-5.6-sol": ["low", "medium", "high", "xhigh", "max"],
+  "gpt-5.6-terra": ["low", "medium", "high", "xhigh", "max"],
+  "google/gemini-3.1-flash-lite": ["low", "medium", "high"],
+  "google/gemini-3.5-flash": ["low", "medium", "high"],
+  "google/gemini-3.5-flash-lite": ["low", "medium", "high"],
+  "google/gemini-3.6-flash": ["low", "medium", "high"],
+  "sakana/fugu-ultra": ["high", "xhigh"],
+  "xai/grok-4.5": ["low", "medium", "high"],
+  "zai-org/GLM-5.2": ["high", "max"],
+}
+
+const PI_THINKING_LEVELS: readonly PiThinkingLevel[] = [
+  "off",
+  "minimal",
+  "low",
+  "medium",
+  "high",
+  "xhigh",
+  "max",
+]
+
+export function thinkingLevelMapForEfforts(
+  efforts: readonly string[],
+): Partial<Record<PiThinkingLevel, string | null>> {
+  const map: Partial<Record<PiThinkingLevel, string | null>> = {}
+  for (const level of PI_THINKING_LEVELS) {
+    if (level === "off") continue
+    map[level] = efforts.includes(level) ? level : null
+  }
+  return map
+}
+
+export interface ThinkingMetadata {
+  thinkingLevelMap: Partial<Record<PiThinkingLevel, string | null>>
+  thinking: {
+    mode: "effort"
+    effortMap: Partial<Record<CommandCodeReasoningEffort, string>>
+    efforts: readonly CommandCodeReasoningEffort[]
+  }
+}
+
+export function thinkingMetadataForModel(modelId: string): ThinkingMetadata | undefined {
+  const efforts = MODEL_EFFORTS[modelId]
+  if (!efforts) return undefined
+  return {
+    thinkingLevelMap: thinkingLevelMapForEfforts(efforts),
+    thinking: {
+      mode: "effort",
+      effortMap: Object.fromEntries(efforts.map((effort) => [effort, effort])),
+      efforts,
+    },
+  }
+}
+
+function isReasoningModel(modelId: string): boolean {
+  return MODEL_EFFORTS[modelId] !== undefined
+}
+
 interface ApiModel {
   id: string
   name: string
@@ -23,6 +167,8 @@ export interface CommandCodeModel {
 interface FetchCommandCodeModelsOptions {
   url?: string
   fetchImpl?: typeof fetch
+  signal?: AbortSignal
+  timeoutMs?: number
 }
 
 interface LoadCommandCodeModelsOptions extends FetchCommandCodeModelsOptions {
@@ -74,10 +220,12 @@ function parseApiModel(value: unknown): ApiModel {
 function parseCachedModel(value: unknown): CommandCodeModel {
   if (!isRecord(value)) throw new Error("Expected cached model entry to be an object")
 
+  const id = stringField(value, "id")
+  booleanField(value, "reasoning")
   return {
-    id: stringField(value, "id"),
+    id,
     name: stringField(value, "name"),
-    reasoning: booleanField(value, "reasoning"),
+    reasoning: isReasoningModel(id),
     contextWindow: positiveNumberField(value, "contextWindow"),
     maxTokens: positiveNumberField(value, "maxTokens"),
   }
@@ -92,6 +240,85 @@ function errorMessage(error: unknown): string {
   return error instanceof Error ? error.message : String(error)
 }
 
+function abortError(reason: unknown): Error {
+  if (reason instanceof Error) return reason
+  return new DOMException("The operation was aborted", "AbortError")
+}
+
+function configuredTimeoutMs(timeoutMs: number | undefined): number {
+  return timeoutMs !== undefined && Number.isFinite(timeoutMs) && timeoutMs > 0
+    ? timeoutMs
+    : DEFAULT_MODELS_TIMEOUT_MS
+}
+
+export function getModelsTimeoutMs(env: NodeJS.ProcessEnv = process.env): number {
+  const raw = env.COMMANDCODE_MODELS_TIMEOUT_MS
+  if (!raw) return DEFAULT_MODELS_TIMEOUT_MS
+
+  const parsed = Number(raw)
+  return configuredTimeoutMs(parsed)
+}
+
+class ModelDiscoveryTimeoutError extends Error {
+  constructor(timeoutMs: number) {
+    super(`Command Code model discovery timed out after ${timeoutMs}ms`)
+    this.name = "ModelDiscoveryTimeoutError"
+  }
+}
+
+function runWithTimeout<T>(
+  operation: (signal: AbortSignal) => Promise<T>,
+  timeoutMs: number,
+  externalSignal: AbortSignal | undefined,
+): Promise<T> {
+  const controller = new AbortController()
+  let timer: ReturnType<typeof setTimeout> | undefined
+  let settled = false
+  let onExternalAbort: (() => void) | undefined
+
+  return new Promise<T>((resolve, reject) => {
+    const cleanup = () => {
+      if (timer !== undefined) clearTimeout(timer)
+      if (onExternalAbort && externalSignal) {
+        externalSignal.removeEventListener("abort", onExternalAbort)
+      }
+    }
+
+    const resolveOnce = (value: T) => {
+      if (settled) return
+      settled = true
+      cleanup()
+      resolve(value)
+    }
+
+    const rejectOnce = (error: unknown) => {
+      if (settled) return
+      settled = true
+      cleanup()
+      reject(error)
+    }
+
+    const abort = (reason: unknown) => {
+      const error = abortError(reason)
+      controller.abort(error)
+      rejectOnce(error)
+    }
+
+    if (externalSignal?.aborted) {
+      abort(externalSignal.reason)
+      return
+    }
+
+    onExternalAbort = () => abort(externalSignal?.reason)
+    externalSignal?.addEventListener("abort", onExternalAbort, { once: true })
+    timer = setTimeout(() => abort(new ModelDiscoveryTimeoutError(timeoutMs)), timeoutMs)
+
+    Promise.resolve()
+      .then(() => operation(controller.signal))
+      .then(resolveOnce, rejectOnce)
+  })
+}
+
 export function commandCodeModelsFromApiResponse(value: unknown): readonly CommandCodeModel[] {
   if (!isRecord(value)) throw new Error("Expected models response to be an object")
   if (value.object !== "list") throw new Error("Expected models response object to be 'list'")
@@ -102,7 +329,7 @@ export function commandCodeModelsFromApiResponse(value: unknown): readonly Comma
   return data.map(parseApiModel).map((model) => ({
     id: model.id,
     name: `${model.name} (CC)`,
-    reasoning: true,
+    reasoning: isReasoningModel(model.id),
     contextWindow: model.contextLength,
     maxTokens: Math.min(model.contextLength, DEFAULT_MAX_OUTPUT_TOKENS),
   }))
@@ -123,19 +350,26 @@ export async function fetchCommandCodeModels(
 ): Promise<readonly CommandCodeModel[]> {
   const url = options.url ?? DEFAULT_MODELS_URL
   const fetchImpl = options.fetchImpl ?? fetch
-  const response = await fetchImpl(url, {
-    headers: {
-      accept: "application/json",
-    },
-  })
+  const body: unknown = await runWithTimeout(
+    async (signal) => {
+      const response = await fetchImpl(url, {
+        headers: {
+          accept: "application/json",
+        },
+        signal,
+      })
 
-  if (!response.ok) {
-    throw new Error(
-      `Failed to fetch Command Code models: ${response.status} ${response.statusText}`,
-    )
-  }
+      if (!response.ok) {
+        throw new Error(
+          `Failed to fetch Command Code models: ${response.status} ${response.statusText}`,
+        )
+      }
 
-  const body: unknown = await response.json()
+      return await response.json()
+    },
+    configuredTimeoutMs(options.timeoutMs),
+    options.signal,
+  )
   return requireModels(commandCodeModelsFromApiResponse(body))
 }
 
@@ -187,6 +421,8 @@ export async function loadCommandCodeModels(
       }
     }
   } catch (liveError) {
+    if (options.signal?.aborted) throw abortError(options.signal.reason ?? liveError)
+
     try {
       const models = await readCommandCodeModelsCache(cachePath)
       return {
@@ -198,7 +434,7 @@ export async function loadCommandCodeModels(
       return {
         models: [],
         source: "empty",
-        warning: `Could not refresh the Command Code model catalog (${errorMessage(liveError)}), and no valid cached catalog is available at ${cachePath} (${errorMessage(cacheError)}). Command Code models will remain unavailable until /reload succeeds.`,
+        warning: `Could not refresh the Command Code model catalog (${errorMessage(liveError)}), and no valid cached catalog is available at ${cachePath} (${errorMessage(cacheError)}). Command Code models will remain unavailable until /commandcode-refresh succeeds.`,
       }
     }
   }

package/src/overflow.ts

@@ -0,0 +1,120 @@
+const COMMAND_CODE_PROVIDER = "commandcode"
+const CONTEXT_OVERFLOW_PREFIX = "context_length_exceeded:"
+
+const COMMAND_CODE_OVERFLOW_PATTERNS = [
+  /\b(?:context[_\s-]*(?:length|window)|model[_\s-]*context[_\s-]*window)[_\s-]*(?:exceeded|overflow(?:ed)?|too[_\s-]*(?:large|long))\b/i,
+  /\b(?:context|prompt|input)[_\s-]*(?:length|window|size|tokens?|limit|maximum)\b[\s\S]{0,120}\b(?:exceed(?:ed|s)?|overflow(?:ed|s)?|too\s+(?:large|long)|(?:maximum|limit)\s+(?:reached|exceeded|hit))\b/i,
+  /\b(?:exceed(?:ed|s)?|overflow(?:ed|s)?|too\s+(?:large|long))\b[\s\S]{0,120}\b(?:context|prompt|input)[_\s-]*(?:length|window|size|tokens?|limit|maximum)\b/i,
+  /\b(?:prompt|input|context)\b[\s\S]{0,32}\btoo\s+(?:large|long)\b/i,
+  /\b(?:prompt|input)[_\s-]*too[_\s-]*(?:large|long)\b/i,
+  /\b(?:prompt|input)[_\s-]*tokens?[_\s-]*(?:limit|maximum|max)[_\s-]*(?:exceeded|reached)\b/i,
+  /\b(?:prompt|input)[_\s-]*(?:tokens?|length|size)\b[\s\S]{0,120}\b(?:limit|maximum)\b[\s\S]{0,40}\b(?:exceed(?:ed|s)?|reached|hit)\b/i,
+  /\b(?:maximum|limit)[_\s-]+(?:allowed[_\s-]+)?(?:context|prompt|input)[_\s-]*(?:length|window|size|tokens?)\b/i,
+]
+
+const NON_OVERFLOW_PATTERNS = [
+  /\brate[_\s-]*limit\b/i,
+  /\btoo\s+many\s+requests\b/i,
+  /\b(?:capacity|quota|throttl(?:e|ed|ing)?|concurren(?:cy|t)|overloaded)\b/i,
+  /\b(?:service|temporarily)\s+unavailable\b/i,
+  /\bstatus(?:[_\s-]*code)?\s*[:=]\s*429\b/i,
+]
+
+const CONTEXT_OVERFLOW_PREFIX_PATTERN = /context_length_exceeded/i
+
+const HTTP_RATE_LIMIT_STATUS_PATTERNS = [
+  /\b(?:api\s+error|http|status(?:[_\s-]*code)?|status[_\s-]*code)\s*[:(]?\s*429\b/i,
+  /["']?(?:status|status[_\s-]*code)["']?\s*:\s*429\b/i,
+]
+
+const BEARER_PATTERN = /\bBearer\s+[A-Za-z0-9._~+/=-]+/gi
+const CREDENTIAL_PATTERN =
+  /\b(?:api[-_ ]?key|apikey|access[-_ ]?token|refresh[-_ ]?token|token|secret|password|authorization)\s*[=:]\s*[^\s,;)]+/gi
+const USER_TOKEN_PATTERN = /\b(?:user|cc)_[A-Za-z0-9_-]{8,}\b/gi
+const QUERY_SECRET_PATTERN =
+  /([?&](?:api[-_ ]?key|apikey|access_token|refresh_token|token|secret|password)=)[^&#\s]+/gi
+const STANDALONE_SECRET_PATTERN =
+  /\b(?:sk|rk|ghp|github_pat|xox[baprs])[-_A-Za-z0-9]{16,}\b|\beyJ[A-Za-z0-9_-]{20,}\.[A-Za-z0-9_-]{10,}\.[A-Za-z0-9_-]{10,}\b/g
+
+export function redactCommandCodeErrorText(value: string): string {
+  return value
+    .replace(BEARER_PATTERN, "Bearer [redacted]")
+    .replace(CREDENTIAL_PATTERN, (match) => {
+      const separatorIndex = match.search(/[=:]/)
+      return separatorIndex < 0 ? "[redacted]" : `${match.slice(0, separatorIndex + 1)}[redacted]`
+    })
+    .replace(USER_TOKEN_PATTERN, "[redacted]")
+    .replace(QUERY_SECRET_PATTERN, "$1[redacted]")
+    .replace(STANDALONE_SECRET_PATTERN, "[redacted]")
+}
+
+function isRecord(value: unknown): value is Record<string, unknown> {
+  return typeof value === "object" && value !== null
+}
+
+export interface CommandCodeMessageLike {
+  role: string
+  provider: string
+  stopReason: string
+  errorMessage?: string
+}
+
+export function commandCodeErrorMessage(value: unknown): string | undefined {
+  if (typeof value === "string") return value
+  if (!isRecord(value)) return undefined
+
+  const record = value
+  const parts: string[] = []
+  for (const key of [
+    "message",
+    "errorMessage",
+    "error",
+    "detail",
+    "details",
+    "code",
+    "type",
+    "reason",
+  ]) {
+    const part = commandCodeErrorMessage(record[key])
+    if (part && !parts.includes(part)) parts.push(part)
+  }
+
+  for (const key of ["status", "statusCode", "httpStatus"]) {
+    const status = record[key]
+    if (typeof status === "string" || typeof status === "number") {
+      const statusPart = `status: ${status}`
+      if (!parts.includes(statusPart)) parts.push(statusPart)
+    }
+  }
+
+  return parts.length > 0 ? redactCommandCodeErrorText(parts.join(": ")) : undefined
+}
+
+export function normalizeCommandCodeErrorMessage(
+  errorMessage: string | undefined,
+): string | undefined {
+  if (!errorMessage) return undefined
+  if (CONTEXT_OVERFLOW_PREFIX_PATTERN.test(errorMessage)) return undefined
+  if (NON_OVERFLOW_PATTERNS.some((pattern) => pattern.test(errorMessage))) return undefined
+  if (HTTP_RATE_LIMIT_STATUS_PATTERNS.some((pattern) => pattern.test(errorMessage)))
+    return undefined
+  if (!COMMAND_CODE_OVERFLOW_PATTERNS.some((pattern) => pattern.test(errorMessage)))
+    return undefined
+
+  return `${CONTEXT_OVERFLOW_PREFIX} ${errorMessage}`
+}
+
+export function normalizeCommandCodeMessage<T extends CommandCodeMessageLike>(
+  message: T,
+  modelProvider?: string,
+): { message: T & { errorMessage: string } } | undefined {
+  if (message.role !== "assistant" || message.stopReason !== "error") return undefined
+  if (message.provider !== COMMAND_CODE_PROVIDER && modelProvider !== COMMAND_CODE_PROVIDER) {
+    return undefined
+  }
+
+  const errorMessage = normalizeCommandCodeErrorMessage(message.errorMessage)
+  if (!errorMessage) return undefined
+
+  return { message: { ...message, errorMessage } }
+}

package/src/pricing.ts

@@ -0,0 +1,226 @@
+export interface CommandCodeModelCostRates {
+  input: number
+  output: number
+  cacheRead: number
+  cacheWrite: number
+}
+
+export interface CommandCodeModelCostTier extends CommandCodeModelCostRates {
+  inputTokensAbove: number
+}
+
+export interface CommandCodeModelCost extends CommandCodeModelCostRates {
+  tiers?: readonly CommandCodeModelCostTier[]
+}
+
+export interface TemporaryPricing {
+  models: readonly string[]
+  expiresOn: string
+  description: string
+}
+
+export const PRICING_SOURCE_URL = "https://commandcode.ai/docs/resources/pricing-limits"
+export const PRICING_LAST_VERIFIED = "2026-08-04"
+
+export const ZERO_MODEL_COST: CommandCodeModelCost = {
+  input: 0,
+  output: 0,
+  cacheRead: 0,
+  cacheWrite: 0,
+}
+
+/**
+ * Display prices in USD per million tokens.
+ *
+ * Context-dependent rates use pi's request-wide input pricing tiers. The
+ * highest threshold exceeded by input + cache reads + cache writes applies to
+ * the full request. The Command Code usage page remains authoritative for the
+ * amount billed for an individual request.
+ */
+export const MODEL_COSTS: Readonly<Record<string, CommandCodeModelCost>> = {
+  // Free models
+  "poolside/laguna-s-2.1-free": { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+  "inclusionai/ling-3.0-flash-free": { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 },
+
+  // Open and open-weight models
+  "tencent/hy3-paid": { input: 0.14, output: 0.58, cacheRead: 0.035, cacheWrite: 0 },
+  "moonshotai/Kimi-K3": { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 0 },
+  "moonshotai/Kimi-K2.7-Code": { input: 0.95, output: 4, cacheRead: 0.19, cacheWrite: 0 },
+  "moonshotai/Kimi-K2.7-Code-Highspeed": {
+    input: 1.9,
+    output: 8,
+    cacheRead: 0.38,
+    cacheWrite: 0,
+  },
+  "moonshotai/Kimi-K2.6": { input: 0.95, output: 4, cacheRead: 0.16, cacheWrite: 0 },
+  "moonshotai/Kimi-K2.5": { input: 0.6, output: 3, cacheRead: 0.1, cacheWrite: 0 },
+  "zai-org/GLM-5.2": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 },
+  "zai-org/GLM-5.2-Fast": { input: 3, output: 10.25, cacheRead: 0.5, cacheWrite: 0 },
+  "zai-org/GLM-5.1": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 },
+  "zai-org/GLM-5": { input: 1, output: 3.2, cacheRead: 0.2, cacheWrite: 0 },
+  "MiniMaxAI/MiniMax-M3": { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0 },
+  "MiniMaxAI/MiniMax-M2.7": { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0 },
+  "MiniMaxAI/MiniMax-M2.5": { input: 0.3, output: 1.2, cacheRead: 0.03, cacheWrite: 0 },
+  // Permanent 75% discount.
+  "deepseek/deepseek-v4-pro": {
+    input: 0.435,
+    output: 0.87,
+    cacheRead: 0.003625,
+    cacheWrite: 0,
+  },
+  "deepseek/deepseek-v4-flash": {
+    input: 0.14,
+    output: 0.28,
+    cacheRead: 0.0028,
+    cacheWrite: 0,
+  },
+  "Qwen/Qwen3.8-Max": { input: 2, output: 6, cacheRead: 0.25, cacheWrite: 2.5 },
+  "Qwen/Qwen3.7-Max": { input: 2.5, output: 7.5, cacheRead: 0.5, cacheWrite: 3.13 },
+  "Qwen/Qwen3.7-Plus": {
+    input: 0.4,
+    output: 1.6,
+    cacheRead: 0.08,
+    cacheWrite: 0.5,
+    tiers: [
+      {
+        inputTokensAbove: 256_000,
+        input: 1.2,
+        output: 4.8,
+        cacheRead: 0.24,
+        cacheWrite: 1.5,
+      },
+    ],
+  },
+  "Qwen/Qwen3.7-Flash": {
+    input: 0.03,
+    output: 0.13,
+    cacheRead: 0.006,
+    cacheWrite: 0.038,
+    tiers: [
+      {
+        inputTokensAbove: 32_000,
+        input: 0.1,
+        output: 0.4,
+        cacheRead: 0.02,
+        cacheWrite: 0.125,
+      },
+      {
+        inputTokensAbove: 256_000,
+        input: 0.2,
+        output: 0.8,
+        cacheRead: 0.04,
+        cacheWrite: 0.25,
+      },
+    ],
+  },
+  "Qwen/Qwen3.6-Max-Preview": {
+    input: 1.3,
+    output: 7.8,
+    cacheRead: 0.26,
+    cacheWrite: 1.63,
+  },
+  "Qwen/Qwen3.6-Plus": { input: 0.5, output: 3, cacheRead: 0.1, cacheWrite: 0 },
+  "stepfun/Step-3.7-Flash": { input: 0.2, output: 1.15, cacheRead: 0.04, cacheWrite: 0 },
+  "stepfun/Step-3.5-Flash": { input: 0.1, output: 0.3, cacheRead: 0.02, cacheWrite: 0 },
+  // Permanent discounted rates.
+  "xiaomi/mimo-v2.5-pro": { input: 0.435, output: 0.87, cacheRead: 0.0036, cacheWrite: 0 },
+  "xiaomi/mimo-v2.5": { input: 0.14, output: 0.28, cacheRead: 0.0028, cacheWrite: 0 },
+  "nvidia/nemotron-3-ultra-550b-a55b": {
+    input: 0.6,
+    output: 2.4,
+    cacheRead: 0.12,
+    cacheWrite: 0,
+  },
+  "sakana/fugu-ultra": { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 0 },
+  "thinkingmachines/inkling": { input: 1, output: 4.05, cacheRead: 0.17, cacheWrite: 0 },
+  "thinkingmachines/inkling-small": {
+    input: 0.5,
+    output: 1.2,
+    cacheRead: 0.1,
+    cacheWrite: 0,
+  },
+  "meta/muse-spark-1.1": { input: 1.25, output: 4.25, cacheRead: 0.15, cacheWrite: 0 },
+
+  // Anthropic
+  // Introductory pricing through 2026-08-31.
+  "claude-sonnet-5": { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2.5 },
+  "claude-sonnet-4-6": { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3.75 },
+  "claude-fable-5": { input: 10, output: 50, cacheRead: 1, cacheWrite: 12.5 },
+  "claude-opus-5": { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 },
+  "claude-opus-4-8": { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 },
+  "claude-opus-4-7": { input: 5, output: 25, cacheRead: 0.5, cacheWrite: 6.25 },
+  "claude-haiku-4-5-20251001": {
+    input: 1,
+    output: 5,
+    cacheRead: 0.1,
+    cacheWrite: 1.25,
+  },
+
+  // OpenAI
+  "gpt-5.6-sol": { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 6.25 },
+  // Discounted rates through 2026-08-14.
+  "gpt-5.6-terra": {
+    input: 1,
+    output: 6,
+    cacheRead: 0.1,
+    cacheWrite: 1.25,
+    tiers: [
+      {
+        inputTokensAbove: 272_000,
+        input: 2,
+        output: 9,
+        cacheRead: 0.2,
+        cacheWrite: 2.5,
+      },
+    ],
+  },
+  "gpt-5.6-luna": {
+    input: 0.1,
+    output: 0.6,
+    cacheRead: 0.01,
+    cacheWrite: 0.125,
+    tiers: [
+      {
+        inputTokensAbove: 272_000,
+        input: 0.2,
+        output: 0.9,
+        cacheRead: 0.02,
+        cacheWrite: 0.25,
+      },
+    ],
+  },
+  "gpt-5.5": { input: 5, output: 30, cacheRead: 0.5, cacheWrite: 0 },
+  "gpt-5.4": { input: 2.5, output: 15, cacheRead: 0.25, cacheWrite: 0 },
+  "gpt-5.3-codex": { input: 2, output: 8, cacheRead: 0.5, cacheWrite: 0 },
+  "gpt-5.4-mini": { input: 0.75, output: 4.5, cacheRead: 0.075, cacheWrite: 0 },
+
+  // Google and xAI
+  "google/gemini-3.6-flash": { input: 1.5, output: 7.5, cacheRead: 0.15, cacheWrite: 0 },
+  "google/gemini-3.5-flash": { input: 1.5, output: 9, cacheRead: 0.15, cacheWrite: 0 },
+  "google/gemini-3.5-flash-lite": {
+    input: 0.3,
+    output: 2.5,
+    cacheRead: 0.03,
+    cacheWrite: 0,
+  },
+  "google/gemini-3.1-flash-lite": {
+    input: 0.25,
+    output: 1.5,
+    cacheRead: 0.03,
+    cacheWrite: 0,
+  },
+  "xai/grok-4.5": { input: 2, output: 6, cacheRead: 0.5, cacheWrite: 0 },
+}
+
+export const TEMPORARY_PRICING: readonly TemporaryPricing[] = [
+  {
+    models: ["gpt-5.6-terra", "gpt-5.6-luna"],
+    expiresOn: "2026-08-14",
+    description: "50% promotional rates",
+  },
+  {
+    models: ["claude-sonnet-5"],
+    expiresOn: "2026-08-31",
+    description: "introductory pricing",
+  },
+]