@dan-ai-studio/dshopencodego 0.1.88 → 0.1.89

package/lib/index.js

@@ -342,11 +342,45 @@ var init_resolve = __esm({
   }
 });
 
+// node_modules/@earendil-works/pi-ai/dist/utils/text.js
+function contentText(content, separator = "\n") {
+  if (typeof content === "string")
+    return content;
+  return content.filter((block) => block.type === "text").map((block) => block.text).join(separator);
+}
+function getSystemMessageText(message) {
+  const parts = [contentText(message.content)];
+  for (const text2 of Object.values(message.sections ?? {})) {
+    if (text2 !== null)
+      parts.push(text2);
+  }
+  return parts.filter((part) => part.length > 0).join("\n\n");
+}
+function renderSystemMessageUpdate(message) {
+  const parts = [];
+  const text2 = contentText(message.content);
+  if (text2.length > 0)
+    parts.push(text2);
+  for (const [name2, value] of Object.entries(message.sections ?? {})) {
+    parts.push(value === null ? `Removed system prompt section "${name2}".` : `Updated system prompt section "${name2}":
+
+${value}`);
+  }
+  return parts.join("\n\n");
+}
+function ownedSlice(text2, start, end = text2.length) {
+  return `${text2.slice(start, end)} `.slice(0, -1);
+}
+var init_text = __esm({
+  "node_modules/@earendil-works/pi-ai/dist/utils/text.js"() {
+  }
+});
+
 // node_modules/@earendil-works/pi-ai/dist/utils/version.js
 var PI_AI_VERSION;
 var init_version = __esm({
   "node_modules/@earendil-works/pi-ai/dist/utils/version.js"() {
-    PI_AI_VERSION = "1.4.0";
+    PI_AI_VERSION = "1.5.0";
   }
 });
 
@@ -454,37 +488,6 @@ var init_model_operations = __esm({
   }
 });
 
-// node_modules/@earendil-works/pi-ai/dist/utils/text.js
-function contentText(content, separator = "\n") {
-  if (typeof content === "string")
-    return content;
-  return content.filter((block) => block.type === "text").map((block) => block.text).join(separator);
-}
-function getSystemMessageText(message) {
-  const parts = [contentText(message.content)];
-  for (const text2 of Object.values(message.sections ?? {})) {
-    if (text2 !== null)
-      parts.push(text2);
-  }
-  return parts.filter((part) => part.length > 0).join("\n\n");
-}
-function renderSystemMessageUpdate(message) {
-  const parts = [];
-  const text2 = contentText(message.content);
-  if (text2.length > 0)
-    parts.push(text2);
-  for (const [name2, value] of Object.entries(message.sections ?? {})) {
-    parts.push(value === null ? `Removed system prompt section "${name2}".` : `Updated system prompt section "${name2}":
-
-${value}`);
-  }
-  return parts.join("\n\n");
-}
-var init_text = __esm({
-  "node_modules/@earendil-works/pi-ai/dist/utils/text.js"() {
-  }
-});
-
 // node_modules/@earendil-works/pi-ai/dist/utils/transcript.js
 function createInitialSystemMessage(systemPrompt, tools) {
   const hasSystemPrompt = systemPrompt !== void 0 && systemPrompt.length > 0;
@@ -731,6 +734,7 @@ function calculateCost(model, usage) {
   usage.cost.cacheRead = rates2.cacheRead / 1e6 * usage.cacheRead;
   usage.cost.cacheWrite = (rates2.cacheWrite * shortWrite + rates2.input * 2 * longWrite) / 1e6;
   usage.cost.total = usage.cost.input + usage.cost.output + usage.cost.cacheRead + usage.cost.cacheWrite;
+  usage.cost.totalFromProvider = false;
   return usage.cost;
 }
 function getSupportedThinkingLevels(model) {
@@ -18293,26 +18297,6 @@ ${json}`
     };
   }
 }
-function parseStreamingJson(partialJson) {
-  if (!partialJson || partialJson.trim() === "") {
-    return {};
-  }
-  try {
-    return parseJsonWithRepair(partialJson);
-  } catch {
-    try {
-      const result = (0, import_partial_json.parse)(partialJson);
-      return result ?? {};
-    } catch {
-      try {
-        const result = (0, import_partial_json.parse)(repairJson(partialJson));
-        return result ?? {};
-      } catch {
-        return {};
-      }
-    }
-  }
-}
 var import_partial_json, VALID_JSON_ESCAPES;
 var init_json_parse = __esm({
   "node_modules/@earendil-works/pi-ai/dist/utils/json-parse.js"() {
@@ -18321,6 +18305,43 @@ var init_json_parse = __esm({
   }
 });
 
+// node_modules/@earendil-works/pi-ai/dist/utils/openrouter-cost.js
+function isValidCost(value) {
+  return typeof value === "number" && Number.isFinite(value) && value >= 0;
+}
+function isOpenRouterModel(model) {
+  return model.provider === "openrouter" || model.baseUrl.includes("openrouter.ai");
+}
+function applyOpenRouterReportedCost(usage, reported) {
+  if (!isValidCost(reported.cost))
+    return;
+  if (reported.is_byok === false) {
+    usage.cost.total = reported.cost;
+    usage.cost.totalFromProvider = true;
+    return;
+  }
+  if (reported.is_byok === true) {
+    const details = reported.cost_details;
+    if (!details)
+      return;
+    let upstreamCost = details.upstream_inference_cost;
+    if (upstreamCost === null) {
+      if (!isValidCost(details.upstream_inference_prompt_cost) || !isValidCost(details.upstream_inference_completions_cost)) {
+        return;
+      }
+      upstreamCost = details.upstream_inference_prompt_cost + details.upstream_inference_completions_cost;
+    }
+    if (!isValidCost(upstreamCost))
+      return;
+    usage.cost.total = reported.cost + upstreamCost;
+    usage.cost.totalFromProvider = true;
+  }
+}
+var init_openrouter_cost = __esm({
+  "node_modules/@earendil-works/pi-ai/dist/utils/openrouter-cost.js"() {
+  }
+});
+
 // node_modules/@earendil-works/pi-ai/dist/utils/provider-retry.js
 function isProviderError(error) {
   if (!(error instanceof Error) || !("status" in error) || !("headers" in error))
@@ -18418,6 +18439,451 @@ var init_sanitize_unicode = __esm({
   }
 });
 
+// node_modules/@earendil-works/pi-ai/dist/utils/streaming-json.js
+function isDelimiter(char) {
+  return char === " " || char === "	" || char === "\n" || char === "\r" || char === "," || char === ":" || char === "{" || char === "}" || char === "[" || char === "]" || char === '"';
+}
+function isHexDigit(char) {
+  return char >= "0" && char <= "9" || char >= "a" && char <= "f" || char >= "A" && char <= "F";
+}
+function scalarValue(token) {
+  for (const [literal, value] of LITERALS)
+    if (token === literal)
+      return value;
+  return NUMBER.test(token) ? Number(token) : void 0;
+}
+function pendingScalarValue(token) {
+  for (const [literal, value] of LITERALS)
+    if (literal.startsWith(token))
+      return value;
+  if (NUMBER.test(token))
+    return Number(token);
+  const mantissa = NUMBER_WITHOUT_EXPONENT_DIGITS.exec(token);
+  return mantissa === null ? void 0 : Number(mantissa[1]);
+}
+function setSlot(container, slot, value) {
+  if (slot.kind === "element")
+    container[slot.index] = value;
+  else if (slot.kind === "member")
+    setMember(container, slot.key, value);
+}
+function setMember(members, key, value) {
+  if (key === "__proto__") {
+    Object.defineProperty(members, key, { value, writable: true, enumerable: true, configurable: true });
+  } else {
+    members[key] = value;
+  }
+}
+var MAX_REPORTED_DEPTH, ROOT_SLOT, CREDIT_PER_APPEND, CREDIT_PER_CHARACTER, MAX_PENDING_SCALAR_LENGTH, NUMBER, NUMBER_WITHOUT_EXPONENT_DIGITS, LITERALS, QUOTE, BACKSLASH, StreamingJsonParser, ToolCallArgumentsStream, SIMPLE_ESCAPES;
+var init_streaming_json = __esm({
+  "node_modules/@earendil-works/pi-ai/dist/utils/streaming-json.js"() {
+    init_json_parse();
+    MAX_REPORTED_DEPTH = 32;
+    ROOT_SLOT = { kind: "root" };
+    CREDIT_PER_APPEND = 256;
+    CREDIT_PER_CHARACTER = 8;
+    MAX_PENDING_SCALAR_LENGTH = 64;
+    NUMBER = /^-?(?:0|[1-9]\d*)(?:\.\d+)?(?:[eE][+-]?\d+)?$/;
+    NUMBER_WITHOUT_EXPONENT_DIGITS = /^(-?(?:0|[1-9]\d*)(?:\.\d+)?)[eE][+-]?$/;
+    LITERALS = [
+      ["true", true],
+      ["false", false],
+      ["null", null]
+    ];
+    QUOTE = 34;
+    BACKSLASH = 92;
+    StreamingJsonParser = class {
+      #root = void 0;
+      #done = false;
+      #stack = [];
+      #string = void 0;
+      #scalar = "";
+      #scalarSlot = void 0;
+      /** Sum of `size + 1` over the open path: what a fresh snapshot copies. */
+      #pathCost = 0;
+      #credit = 0;
+      #dirty = false;
+      #snapshot = void 0;
+      /** Changes of the piece being read. */
+      #changes = [];
+      /**
+       * Read the next piece of text; returns the changes it made, in order. Costs time linear in `text.length`.
+       */
+      append(text2) {
+        this.#credit += CREDIT_PER_APPEND + CREDIT_PER_CHARACTER * text2.length;
+        this.#changes = [];
+        if (this.#string !== void 0)
+          this.#string.change = void 0;
+        let index = 0;
+        while (index < text2.length && !this.#done) {
+          const open = this.#string;
+          if (open !== void 0)
+            index = this.#readString(open, text2, index);
+          else if (this.#scalar !== "")
+            index = this.#readScalar(text2, index);
+          else
+            this.#structural(text2[index++]);
+        }
+        return this.#changes;
+      }
+      /**
+       * The value of the text read so far: `undefined` before a value starts, otherwise a plain JSON value that never
+       * changes afterwards. Returns the previous snapshot while the text read since did not change the value, or while a
+       * fresh copy of the open path is not yet paid for (see the module comment).
+       */
+      snapshot() {
+        if (!this.#dirty)
+          return this.#snapshot;
+        const pending = this.#pendingScalar();
+        const cost = this.#pathCost + (pending === void 0 ? 0 : 1);
+        if (cost > this.#credit)
+          return this.#snapshot;
+        this.#credit -= cost;
+        this.#dirty = false;
+        this.#snapshot = this.#copyOpenPath(pending);
+        return this.#snapshot;
+      }
+      /** Read string characters from `start` until the string ends or the text does; returns the next index. */
+      #readString(open, text2, start) {
+        let runStart = start;
+        for (let index = start; index < text2.length; index++) {
+          const code = text2.charCodeAt(index);
+          if (open.escape !== "") {
+            if (this.#continueEscape(open, text2[index])) {
+              runStart = index + 1;
+              continue;
+            }
+            runStart = index;
+          }
+          if (code === QUOTE) {
+            if (index > runStart)
+              this.#appendString(open, text2.slice(runStart, index));
+            this.#endString(open);
+            return index + 1;
+          }
+          if (code === BACKSLASH) {
+            if (index > runStart)
+              this.#appendString(open, text2.slice(runStart, index));
+            open.escape = "\\";
+            runStart = index + 1;
+          }
+        }
+        if (text2.length > runStart)
+          this.#appendString(open, text2.slice(runStart));
+        return text2.length;
+      }
+      /** Read number or literal characters from `start` up to the next delimiter; returns the delimiter's index. */
+      #readScalar(text2, start) {
+        let end = start;
+        while (end < text2.length && !isDelimiter(text2[end]))
+          end++;
+        if (end > start) {
+          this.#scalar += text2.slice(start, end);
+          if (this.#scalarSlot !== void 0)
+            this.#dirty = true;
+        }
+        if (end < text2.length)
+          this.#endScalar();
+        return end;
+      }
+      /** Handle one character outside strings. */
+      #structural(char) {
+        switch (char) {
+          case " ":
+          case "	":
+          case "\n":
+          case "\r":
+            return;
+          case '"':
+            this.#startString();
+            return;
+          case "{":
+          case "[":
+            this.#openContainer(char === "{" ? {} : []);
+            return;
+          case "}":
+          case "]":
+            this.#closeContainer();
+            return;
+          case ":": {
+            const frame = this.#stack.at(-1);
+            if (frame !== void 0 && !Array.isArray(frame.container) && frame.phase === "colon")
+              frame.phase = "value";
+            return;
+          }
+          case ",": {
+            const frame = this.#stack.at(-1);
+            if (frame !== void 0)
+              frame.phase = Array.isArray(frame.container) ? "value" : "key";
+            return;
+          }
+          default:
+            this.#scalar = char;
+            this.#scalarSlot = this.#valueSlot();
+            if (this.#scalarSlot !== void 0)
+              this.#dirty = true;
+        }
+      }
+      /** Where a value starting now goes, or `undefined` when this position takes no value. */
+      #valueSlot() {
+        const frame = this.#stack.at(-1);
+        if (frame === void 0)
+          return this.#root === void 0 ? ROOT_SLOT : void 0;
+        if (Array.isArray(frame.container))
+          return { kind: "element", index: frame.container.length };
+        if ((frame.phase === "value" || frame.phase === "colon") && frame.key !== void 0) {
+          return { kind: "member", key: frame.key };
+        }
+        return void 0;
+      }
+      #startString() {
+        const frame = this.#stack.at(-1);
+        if (frame !== void 0 && !Array.isArray(frame.container) && (frame.phase === "key" || frame.phase === "comma")) {
+          this.#string = { target: "key", slot: void 0, text: "", escape: "", path: void 0, change: void 0 };
+          return;
+        }
+        const slot = this.#valueSlot();
+        const path5 = this.#reportedPath(slot);
+        this.#string = { target: "value", slot, text: "", escape: "", path: path5, change: void 0 };
+        this.#place(slot, "");
+        if (path5 !== void 0)
+          this.#string.change = this.#report({ kind: "set", path: path5, value: "" });
+      }
+      /** The path of a value at `slot` in the innermost open container, when its changes are reported. */
+      #reportedPath(slot) {
+        if (slot === void 0)
+          return void 0;
+        if (slot.kind === "root")
+          return [];
+        const frame = this.#stack.at(-1);
+        if (!frame.reportsInside)
+          return void 0;
+        return [...frame.path, slot.kind === "member" ? slot.key : slot.index];
+      }
+      #report(change) {
+        this.#changes.push(change);
+        return change;
+      }
+      /** Continue an escape with `char`; returns false when `char` is not part of it and must be read normally. */
+      #continueEscape(open, char) {
+        if (open.escape === "\\") {
+          const decoded = SIMPLE_ESCAPES[char];
+          if (decoded !== void 0) {
+            open.escape = "";
+            this.#appendString(open, decoded);
+            return true;
+          }
+          if (char === "u") {
+            open.escape = "\\u";
+            return true;
+          }
+          open.escape = "";
+          this.#appendString(open, "\\");
+          return false;
+        }
+        if (isHexDigit(char)) {
+          open.escape += char;
+          if (open.escape.length < 6)
+            return true;
+          const decoded = String.fromCharCode(Number.parseInt(open.escape.slice(2), 16));
+          open.escape = "";
+          this.#appendString(open, decoded);
+          return true;
+        }
+        const literal = open.escape;
+        open.escape = "";
+        this.#appendString(open, literal);
+        return false;
+      }
+      #appendString(open, text2) {
+        open.text += text2;
+        if (open.target !== "value")
+          return;
+        this.#place(open.slot, open.text);
+        if (open.path === void 0)
+          return;
+        const change = open.change;
+        if (change !== void 0 && this.#changes.at(-1) === change) {
+          if (change.kind === "set")
+            change.value = change.value + text2;
+          else
+            change.text += text2;
+          return;
+        }
+        open.change = this.#report({ kind: "append", path: open.path, text: text2 });
+      }
+      #endString(open) {
+        this.#string = void 0;
+        const frame = this.#stack.at(-1);
+        if (open.target === "key") {
+          if (frame === void 0)
+            return;
+          frame.key = open.text;
+          frame.phase = "colon";
+          return;
+        }
+        this.#valueEnded(frame);
+      }
+      #endScalar() {
+        const token = this.#scalar;
+        const slot = this.#scalarSlot;
+        this.#scalar = "";
+        this.#scalarSlot = void 0;
+        const value = scalarValue(token);
+        if (value !== void 0) {
+          this.#place(slot, value);
+          const path5 = this.#reportedPath(slot);
+          if (path5 !== void 0)
+            this.#report({ kind: "set", path: path5, value });
+        }
+        this.#valueEnded(this.#stack.at(-1));
+      }
+      /** The value of the unterminated scalar at the end of the text, if it shows. */
+      #pendingScalar() {
+        const slot = this.#scalarSlot;
+        const token = this.#scalar;
+        if (slot === void 0 || token === "" || token.length > MAX_PENDING_SCALAR_LENGTH)
+          return void 0;
+        const value = pendingScalarValue(token);
+        return value === void 0 ? void 0 : { slot, value };
+      }
+      #openContainer(container) {
+        const slot = this.#valueSlot();
+        const path5 = this.#reportedPath(slot);
+        this.#place(slot, container);
+        this.#stack.push({
+          container,
+          slot,
+          path: path5,
+          reportsInside: path5 !== void 0 && path5.length < MAX_REPORTED_DEPTH,
+          size: 0,
+          key: void 0,
+          phase: Array.isArray(container) ? "value" : "key"
+        });
+        this.#pathCost += 1;
+        this.#dirty = true;
+        if (path5 !== void 0)
+          this.#report({ kind: "set", path: path5, value: Array.isArray(container) ? [] : {} });
+      }
+      #closeContainer() {
+        const frame = this.#stack.pop();
+        if (frame === void 0)
+          return;
+        this.#pathCost -= frame.size + 1;
+        this.#dirty = true;
+        if (frame.path !== void 0 && !frame.reportsInside) {
+          this.#report({ kind: "set", path: frame.path, value: frame.container });
+        }
+        this.#valueEnded(this.#stack.at(-1));
+      }
+      /** A value in `frame`, or the root value when there is no frame, has ended. */
+      #valueEnded(frame) {
+        if (frame === void 0) {
+          if (this.#root !== void 0)
+            this.#done = true;
+          return;
+        }
+        frame.phase = "comma";
+      }
+      /** Set the value at `slot` in the open path; nothing for a value in an invalid position. */
+      #place(slot, value) {
+        if (slot === void 0)
+          return;
+        this.#dirty = true;
+        if (slot.kind === "root") {
+          this.#root = value;
+          return;
+        }
+        const frame = this.#stack.at(-1);
+        if (slot.kind === "element") {
+          const elements = frame.container;
+          if (slot.index === elements.length) {
+            frame.size++;
+            this.#pathCost++;
+          }
+          elements[slot.index] = value;
+          return;
+        }
+        const members = frame.container;
+        if (!Object.hasOwn(members, slot.key)) {
+          frame.size++;
+          this.#pathCost++;
+        }
+        setMember(members, slot.key, value);
+      }
+      /** A fresh snapshot: copies of the open containers, innermost first, each holding the copy of its open child. */
+      #copyOpenPath(pending) {
+        if (this.#stack.length === 0)
+          return pending?.slot.kind === "root" ? pending.value : this.#root;
+        let child;
+        let childSlot;
+        for (let depth = this.#stack.length - 1; depth >= 0; depth--) {
+          const frame = this.#stack[depth];
+          const copy = Array.isArray(frame.container) ? frame.container.slice() : { ...frame.container };
+          if (depth === this.#stack.length - 1 && pending !== void 0)
+            setSlot(copy, pending.slot, pending.value);
+          if (child !== void 0 && childSlot !== void 0)
+            setSlot(copy, childSlot, child);
+          child = copy;
+          childSlot = frame.slot;
+        }
+        return childSlot?.kind === "root" ? child : this.#root;
+      }
+    };
+    ToolCallArgumentsStream = class {
+      #text = "";
+      #parser = new StreamingJsonParser();
+      #snapshot = void 0;
+      #arguments = {};
+      /** The raw argument text received so far. */
+      get text() {
+        return this.#text;
+      }
+      /** Append a piece of argument text; returns the partial arguments, a new object whenever they changed. */
+      append(delta) {
+        this.#text += delta;
+        this.#parser.append(delta);
+        return this.#partial();
+      }
+      /**
+       * Take the provider's complete argument text. When it continues the text so far, returns the missing piece as
+       * `delta`; otherwise starts over with `text`. Either way returns the partial arguments of `text`.
+       */
+      replace(text2) {
+        if (text2.startsWith(this.#text)) {
+          const delta = text2.slice(this.#text.length);
+          return { delta, arguments: delta === "" ? this.#partial() : this.append(delta) };
+        }
+        this.#text = "";
+        this.#parser = new StreamingJsonParser();
+        return { delta: void 0, arguments: this.append(text2) };
+      }
+      /** The final arguments: one full parse of the complete text, which reports malformed arguments. */
+      finish() {
+        return parseToolCallArguments(this.#text);
+      }
+      #partial() {
+        const snapshot2 = this.#parser.snapshot();
+        if (snapshot2 === this.#snapshot)
+          return this.#arguments;
+        this.#snapshot = snapshot2;
+        this.#arguments = typeof snapshot2 === "object" && snapshot2 !== null && !Array.isArray(snapshot2) ? snapshot2 : {};
+        return this.#arguments;
+      }
+    };
+    SIMPLE_ESCAPES = {
+      '"': '"',
+      "\\": "\\",
+      "/": "/",
+      b: "\b",
+      f: "\f",
+      n: "\n",
+      r: "\r",
+      t: "	"
+    };
+  }
+});
+
 // node_modules/@earendil-works/pi-ai/dist/api/constrained-sampling.js
 function isJsonSchemaObject(value) {
   return typeof value === "object" && value !== null && !Array.isArray(value);
@@ -18511,6 +18977,13 @@ function makeStrictJsonSchema(schema, isUnsupportedKeyword) {
   }
   return cloned;
 }
+function prefersJsonSchemaStrictSampling(tool) {
+  const config = tool.constrainedSampling;
+  return !!config && config.type === "json_schema" && config.strict === "prefer";
+}
+function isStrictToolLimitErrorMessage(message) {
+  return STRICT_TOOL_LIMIT_ERROR_MESSAGES.some((limitMessage) => message.includes(limitMessage));
+}
 function getJsonSchemaToolParameters(tool, strict) {
   return strict === true ? makeStrictJsonSchema(tool.parameters) : tool.parameters;
 }
@@ -18521,30 +18994,35 @@ function getGrammarToolInput(toolName2, arguments_, inputProperty) {
   }
   return input;
 }
-function appendGrammarToolInputJsonDelta(buffer, inputProperty, nextInput, close) {
+function appendGrammarToolInputJsonDelta(buffer, inputProperty, inputDelta) {
   if (buffer.closed) {
-    if (close && nextInput === buffer.input)
-      return void 0;
     throw new Error(`grammar tool input for property "${inputProperty}" changed after it was closed`);
   }
-  if (!nextInput.startsWith(buffer.input)) {
-    throw new Error(`grammar tool input for property "${inputProperty}" changed non-monotonically`);
-  }
-  const inputDelta = nextInput.slice(buffer.input.length);
-  if (!close && inputDelta.length === 0)
+  if (inputDelta.length === 0)
     return void 0;
-  let delta = "";
-  if (!buffer.started) {
-    delta += `{${JSON.stringify(inputProperty)}:"`;
-    buffer.started = true;
+  buffer.input += inputDelta;
+  return grammarToolInputJsonStart(buffer, inputProperty) + JSON.stringify(inputDelta).slice(1, -1);
+}
+function closeGrammarToolInputJson(buffer, inputProperty, finalInput) {
+  if (buffer.closed) {
+    if (finalInput === void 0 || finalInput === buffer.input)
+      return void 0;
+    throw new Error(`grammar tool input for property "${inputProperty}" changed after it was closed`);
   }
-  delta += JSON.stringify(inputDelta).slice(1, -1);
-  buffer.input = nextInput;
-  if (close) {
-    delta += '"}';
-    buffer.closed = true;
+  if (finalInput !== void 0 && !finalInput.startsWith(buffer.input)) {
+    throw new Error(`grammar tool input for property "${inputProperty}" changed non-monotonically`);
   }
-  return delta;
+  const rest = finalInput === void 0 ? "" : finalInput.slice(buffer.input.length);
+  if (finalInput !== void 0)
+    buffer.input = finalInput;
+  buffer.closed = true;
+  return `${grammarToolInputJsonStart(buffer, inputProperty)}${JSON.stringify(rest).slice(1, -1)}"}`;
+}
+function grammarToolInputJsonStart(buffer, inputProperty) {
+  if (buffer.started)
+    return "";
+  buffer.started = true;
+  return `{${JSON.stringify(inputProperty)}:"`;
 }
 function inferGrammarInputProperty(tool) {
   const schema = tool.parameters;
@@ -18620,7 +19098,7 @@ function createGrammarToolInputProperties(tools, supportsOpenAIGrammarTools) {
   }
   return properties;
 }
-var UnsupportedStrictJsonSchemaError, UNSUPPORTED_STRICT_SCHEMA_KEYS;
+var UnsupportedStrictJsonSchemaError, UNSUPPORTED_STRICT_SCHEMA_KEYS, STRICT_TOOL_LIMIT_ERROR_MESSAGES;
 var init_constrained_sampling = __esm({
   "node_modules/@earendil-works/pi-ai/dist/api/constrained-sampling.js"() {
     UnsupportedStrictJsonSchemaError = class extends Error {
@@ -18643,6 +19121,11 @@ var init_constrained_sampling = __esm({
       "then",
       "else"
     ];
+    STRICT_TOOL_LIMIT_ERROR_MESSAGES = [
+      "The compiled grammar is too large",
+      "Too many strict tools (",
+      "Schemas contains too many parameters with union types ("
+    ];
   }
 });
 
@@ -19406,10 +19889,7 @@ function buildParams(model, context, isOAuthToken2, options, disablePreferredStr
   let hasPreferredStrictTools = false;
   const convertToolDefinitions = (tools, toolCacheControl2) => {
     const definitions = convertTools(tools, isOAuthToken2, compat.supportsEagerToolInputStreaming, compat.supportsStrictTools, toolCacheControl2, disablePreferredStrictTools);
-    if (definitions.some((definition, index) => {
-      const config = tools[index].constrainedSampling;
-      return definition.strict === true && config && config.type === "json_schema" && config.strict === "prefer";
-    })) {
+    if (definitions.some((definition, index) => definition.strict === true && prefersJsonSchemaStrictSampling(tools[index]))) {
       hasPreferredStrictTools = true;
     }
     return definitions;
@@ -19472,7 +19952,7 @@ function buildParams(model, context, isOAuthToken2, options, disablePreferredStr
       display: options?.thinkingDisplay ?? "summarized",
       block_binding: { prefix_mismatch_behavior: "drop_block" }
     };
-    params.output_config = { effort: "high" };
+    params.output_config = { effort: "low" };
   } else if (model.reasoning && options?.thinkingEnabled) {
     const display = options.thinkingDisplay ?? "summarized";
     if (model.compat?.forceAdaptiveThinking === true) {
@@ -19583,7 +20063,6 @@ function convertMessages(transformedMessages, isOAuthToken2, cacheControl, allow
         });
       }
     } else if (msg.role === "assistant") {
-      flushPendingSystemMessages();
       const blocks = [];
       for (const block of msg.content) {
         if (block.type === "text") {
@@ -19632,6 +20111,7 @@ function convertMessages(transformedMessages, isOAuthToken2, cacheControl, allow
       }
       if (blocks.length === 0)
         continue;
+      flushPendingSystemMessages();
       const messageIndex = params.length;
       params.push({
         role: "assistant",
@@ -19704,8 +20184,7 @@ function convertTools(tools, isOAuthToken2, supportsEagerToolInputStreaming, sup
   if (!tools)
     return [];
   return tools.map((tool, index) => {
-    const config = tool.constrainedSampling;
-    const strict = disablePreferredStrictTools && config && config.type === "json_schema" && config.strict === "prefer" ? void 0 : resolveJsonSchemaStrictSampling(tool, supportsStrictTools, isAnthropicStrictUnsupportedKeyword);
+    const strict = disablePreferredStrictTools && prefersJsonSchemaStrictSampling(tool) ? void 0 : resolveJsonSchemaStrictSampling(tool, supportsStrictTools, isAnthropicStrictUnsupportedKeyword);
     const parameters = getJsonSchemaToolParameters(tool, strict);
     const schema = parameters;
     const legacyInputSchema = {
@@ -19762,9 +20241,11 @@ var init_anthropic_messages = __esm({
     init_headers2();
     init_identity();
     init_json_parse();
+    init_openrouter_cost();
     init_provider_env();
     init_provider_retry();
     init_sanitize_unicode();
+    init_streaming_json();
     init_text();
     init_transcript();
     init_constrained_sampling();
@@ -19901,11 +20382,7 @@ var init_anthropic_messages = __esm({
               });
               break;
             } catch (error) {
-              if (options?.signal?.aborted || disablePreferredStrictTools || !built.hasPreferredStrictTools || !(error instanceof Anthropic.APIError) || error.status !== 400 || error.type !== "invalid_request_error" || ![
-                "The compiled grammar is too large",
-                "Too many strict tools (",
-                "Schemas contains too many parameters with union types ("
-              ].some((message) => error.message.includes(message))) {
+              if (options?.signal?.aborted || disablePreferredStrictTools || !built.hasPreferredStrictTools || !(error instanceof Anthropic.APIError) || error.status !== 400 || error.type !== "invalid_request_error" || !isStrictToolLimitErrorMessage(error.message)) {
                 throw error;
               }
               disablePreferredStrictTools = true;
@@ -19919,7 +20396,11 @@ var init_anthropic_messages = __esm({
           }
           await options?.onResponse?.({ status: response.status, headers: headersToRecord(response.headers) }, model);
           stream5.push({ type: "start", partial: output });
-          const blocks = output.content;
+          const openBlocks = /* @__PURE__ */ new Map();
+          const openBlock = (anthropicIndex) => {
+            const open = openBlocks.get(anthropicIndex);
+            return open === void 0 ? void 0 : { ...open, block: output.content[open.contentIndex] };
+          };
           for await (const event of iterateAnthropicEvents(response, options?.signal)) {
             await options?.onProviderStreamEvent?.(event, model);
             if (event.type === "message_start") {
@@ -19947,95 +20428,82 @@ var init_anthropic_messages = __esm({
                 continue;
               }
               if (event.content_block.type === "text") {
-                const block = {
-                  type: "text",
-                  text: event.content_block.text ?? "",
-                  index: event.index
-                };
-                output.content.push(block);
+                output.content.push({ type: "text", text: event.content_block.text ?? "" });
+                openBlocks.set(event.index, { contentIndex: output.content.length - 1 });
                 stream5.push({ type: "text_start", contentIndex: output.content.length - 1, partial: output });
               } else if (event.content_block.type === "thinking") {
-                const block = {
+                output.content.push({
                   type: "thinking",
                   thinking: event.content_block.thinking ?? "",
-                  thinkingSignature: event.content_block.signature ?? "",
-                  index: event.index
-                };
-                output.content.push(block);
+                  thinkingSignature: event.content_block.signature ?? ""
+                });
+                openBlocks.set(event.index, { contentIndex: output.content.length - 1 });
                 stream5.push({ type: "thinking_start", contentIndex: output.content.length - 1, partial: output });
               } else if (event.content_block.type === "redacted_thinking") {
-                const block = {
+                output.content.push({
                   type: "thinking",
                   thinking: "[Reasoning redacted]",
                   thinkingSignature: event.content_block.data,
-                  redacted: true,
-                  index: event.index
-                };
-                output.content.push(block);
+                  redacted: true
+                });
+                openBlocks.set(event.index, { contentIndex: output.content.length - 1 });
                 stream5.push({ type: "thinking_start", contentIndex: output.content.length - 1, partial: output });
               } else if (event.content_block.type === "tool_use") {
-                const block = {
+                output.content.push({
                   type: "toolCall",
                   id: event.content_block.id,
                   name: isOAuth ? fromClaudeCodeName(event.content_block.name, currentTools) : event.content_block.name,
-                  arguments: event.content_block.input ?? {},
-                  partialJson: "",
-                  index: event.index
-                };
-                output.content.push(block);
+                  arguments: event.content_block.input ?? {}
+                });
+                openBlocks.set(event.index, {
+                  contentIndex: output.content.length - 1,
+                  arguments: new ToolCallArgumentsStream()
+                });
                 stream5.push({ type: "toolcall_start", contentIndex: output.content.length - 1, partial: output });
               }
             } else if (event.type === "content_block_delta") {
+              const open = openBlock(event.index);
               if (event.delta.type === "text_delta") {
-                const index = blocks.findIndex((b) => b.index === event.index);
-                const block = blocks[index];
-                if (block && block.type === "text") {
-                  block.text += event.delta.text;
+                if (open?.block.type === "text") {
+                  open.block.text += event.delta.text;
                   stream5.push({
                     type: "text_delta",
-                    contentIndex: index,
+                    contentIndex: open.contentIndex,
                     delta: event.delta.text,
                     partial: output
                   });
                 }
               } else if (event.delta.type === "thinking_delta") {
-                const index = blocks.findIndex((b) => b.index === event.index);
-                const block = blocks[index];
-                if (block && block.type === "thinking") {
-                  block.thinking += event.delta.thinking;
+                if (open?.block.type === "thinking") {
+                  open.block.thinking += event.delta.thinking;
                   stream5.push({
                     type: "thinking_delta",
-                    contentIndex: index,
+                    contentIndex: open.contentIndex,
                     delta: event.delta.thinking,
                     partial: output
                   });
                 }
               } else if (event.delta.type === "input_json_delta") {
-                const index = blocks.findIndex((b) => b.index === event.index);
-                const block = blocks[index];
-                if (block && block.type === "toolCall") {
-                  block.partialJson += event.delta.partial_json;
-                  block.arguments = parseStreamingJson(block.partialJson);
+                if (open?.block.type === "toolCall" && open.arguments !== void 0) {
+                  open.block.arguments = open.arguments.append(event.delta.partial_json);
                   stream5.push({
                     type: "toolcall_delta",
-                    contentIndex: index,
+                    contentIndex: open.contentIndex,
                     delta: event.delta.partial_json,
                     partial: output
                   });
                 }
               } else if (event.delta.type === "signature_delta") {
-                const index = blocks.findIndex((b) => b.index === event.index);
-                const block = blocks[index];
-                if (block && block.type === "thinking") {
-                  block.thinkingSignature = block.thinkingSignature || "";
-                  block.thinkingSignature += event.delta.signature;
+                if (open?.block.type === "thinking") {
+                  open.block.thinkingSignature = open.block.thinkingSignature || "";
+                  open.block.thinkingSignature += event.delta.signature;
                 }
               }
             } else if (event.type === "content_block_stop") {
-              const index = blocks.findIndex((b) => b.index === event.index);
-              const block = blocks[index];
-              if (block) {
-                delete block.index;
+              const open = openBlock(event.index);
+              openBlocks.delete(event.index);
+              if (open) {
+                const { block, contentIndex: index } = open;
                 if (block.type === "text") {
                   stream5.push({
                     type: "text_end",
@@ -20050,9 +20518,8 @@ var init_anthropic_messages = __esm({
                     content: block.thinking,
                     partial: output
                   });
-                } else if (block.type === "toolCall") {
-                  Object.assign(block, parseToolCallArguments(block.partialJson));
-                  delete block.partialJson;
+                } else if (block.type === "toolCall" && open.arguments !== void 0) {
+                  Object.assign(block, open.arguments.finish());
                   stream5.push({
                     type: "toolcall_end",
                     contentIndex: index,
@@ -20097,6 +20564,9 @@ var init_anthropic_messages = __esm({
               }
               output.usage.totalTokens = output.usage.input + output.usage.output + output.usage.cacheRead + output.usage.cacheWrite;
               calculateCost(usageModel, output.usage);
+              if (isOpenRouterModel(model) && event.usage) {
+                applyOpenRouterReportedCost(output.usage, event.usage);
+              }
             }
           }
           if (options?.signal?.aborted) {
@@ -20124,10 +20594,6 @@ var init_anthropic_messages = __esm({
           stream5.push({ type: "done", reason: output.stopReason, message: output });
           stream5.end();
         } catch (error) {
-          for (const block of output.content) {
-            delete block.index;
-            delete block.partialJson;
-          }
           output.stopReason = options?.signal?.aborted ? "aborted" : "error";
           output.errorMessage = error instanceof Error ? error.message : JSON.stringify(error);
           stream5.push({ type: "error", reason: output.stopReason, error: output });
@@ -40373,7 +40839,7 @@ function formatProviderError(norm, prefix) {
 function truncateErrorText(text2, maxChars) {
   if (text2.length <= maxChars)
     return text2;
-  return `${text2.slice(0, maxChars)}... [truncated ${text2.length - maxChars} chars]`;
+  return `${ownedSlice(text2, 0, maxChars)}... [truncated ${text2.length - maxChars} chars]`;
 }
 function safeJsonStringify2(value) {
   try {
@@ -40386,6 +40852,7 @@ function safeJsonStringify2(value) {
 var MAX_PROVIDER_ERROR_BODY_CHARS;
 var init_error_body = __esm({
   "node_modules/@earendil-works/pi-ai/dist/utils/error-body.js"() {
+    init_text();
     MAX_PROVIDER_ERROR_BODY_CHARS = 4e3;
   }
 });
@@ -41104,6 +41571,9 @@ function parseChunkUsage(rawUsage, model) {
     cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0, total: 0 }
   };
   calculateCost(model, usage);
+  if (isOpenRouterModel(model)) {
+    applyOpenRouterReportedCost(usage, rawUsage);
+  }
   return usage;
 }
 function mapStopReason2(reason) {
@@ -41221,10 +41691,11 @@ var init_openai_completions = __esm({
     init_hash();
     init_headers2();
     init_identity();
-    init_json_parse();
+    init_openrouter_cost();
     init_provider_env();
     init_provider_retry();
     init_sanitize_unicode();
+    init_streaming_json();
     init_text();
     init_transcript();
     init_constrained_sampling();
@@ -41289,23 +41760,14 @@ var init_openai_completions = __esm({
           let hasFinishReason = false;
           const toolCallBlocksByIndex = /* @__PURE__ */ new Map();
           const toolCallBlocksById = /* @__PURE__ */ new Map();
+          const toolCallStates = /* @__PURE__ */ new Map();
           const blocks = output.content;
           const getContentIndex = (block) => blocks.indexOf(block);
-          const getCustomToolCallInput2 = (block) => {
-            const property2 = block.customInput?.property;
-            if (property2 === void 0)
-              return "";
-            const value = block.arguments[property2];
-            return typeof value === "string" ? value : "";
-          };
-          const appendCustomToolCallInput2 = (block, nextInput, close) => {
-            const customInput = block.customInput;
-            if (!customInput)
-              return void 0;
-            const delta = appendGrammarToolInputJsonDelta(customInput.jsonBuffer, customInput.property, nextInput, close);
-            block.arguments = { [customInput.property]: nextInput };
-            return delta;
-          };
+          const customInput2 = (property2) => ({
+            kind: "custom",
+            property: property2,
+            buffer: { input: "", started: false, closed: false }
+          });
           const finishBlock = (block) => {
             const contentIndex = getContentIndex(block);
             if (contentIndex === -1) {
@@ -41327,8 +41789,10 @@ var init_openai_completions = __esm({
                 partial: output
               });
             } else if (block.type === "toolCall") {
-              if (block.customInput) {
-                const delta = appendCustomToolCallInput2(block, getCustomToolCallInput2(block), true);
+              const input = toolCallStates.get(block).input;
+              toolCallStates.delete(block);
+              if (input.kind === "custom") {
+                const delta = closeGrammarToolInputJson(input.buffer, input.property, void 0);
                 if (delta !== void 0) {
                   stream5.push({
                     type: "toolcall_delta",
@@ -41338,11 +41802,8 @@ var init_openai_completions = __esm({
                   });
                 }
               } else {
-                Object.assign(block, parseToolCallArguments(block.partialArgs));
+                Object.assign(block, input.arguments.finish());
               }
-              delete block.partialArgs;
-              delete block.customInput;
-              delete block.streamIndex;
               stream5.push({
                 type: "toolcall_end",
                 contentIndex,
@@ -41380,16 +41841,16 @@ var init_openai_completions = __esm({
             }
             if (!block) {
               const customInputProperty = toolCall.custom && !toolCall.function ? grammarToolInputProperties.get(name2) ?? "input" : void 0;
-              const hasCustomInput = customInputProperty !== void 0;
               block = {
                 type: "toolCall",
                 id: toolCall.id || "",
                 name: name2,
-                arguments: hasCustomInput ? { [customInputProperty]: "" } : {},
-                partialArgs: hasCustomInput ? void 0 : "",
-                customInput: hasCustomInput ? { property: customInputProperty, jsonBuffer: { input: "", started: false, closed: false } } : void 0,
-                streamIndex
+                arguments: customInputProperty !== void 0 ? { [customInputProperty]: "" } : {}
               };
+              toolCallStates.set(block, {
+                streamIndex,
+                input: customInputProperty !== void 0 ? customInput2(customInputProperty) : { kind: "json", arguments: new ToolCallArgumentsStream() }
+              });
               if (streamIndex !== void 0) {
                 toolCallBlocksByIndex.set(streamIndex, block);
               }
@@ -41403,8 +41864,9 @@ var init_openai_completions = __esm({
                 partial: output
               });
             }
-            if (streamIndex !== void 0 && block.streamIndex === void 0) {
-              block.streamIndex = streamIndex;
+            const state2 = toolCallStates.get(block);
+            if (streamIndex !== void 0 && state2.streamIndex === void 0) {
+              state2.streamIndex = streamIndex;
               toolCallBlocksByIndex.set(streamIndex, block);
             }
             if (toolCall.id) {
@@ -41413,16 +41875,12 @@ var init_openai_completions = __esm({
             if (!block.name && name2) {
               block.name = name2;
             }
-            if (toolCall.custom && !toolCall.function && !block.customInput) {
+            if (toolCall.custom && !toolCall.function && state2.input.kind !== "custom") {
               const customInputProperty = grammarToolInputProperties.get(block.name) ?? "input";
               block.arguments = { [customInputProperty]: "" };
-              block.customInput = {
-                property: customInputProperty,
-                jsonBuffer: { input: "", started: false, closed: false }
-              };
-              delete block.partialArgs;
+              state2.input = customInput2(customInputProperty);
             }
-            return block;
+            return { block, state: state2 };
           };
           for await (const chunk of openaiStream) {
             await options?.onProviderStreamEvent?.(chunk, model);
@@ -41497,7 +41955,7 @@ var init_openai_completions = __esm({
               }
               if (choice?.delta?.tool_calls) {
                 for (const toolCall of choice.delta.tool_calls) {
-                  const block = ensureToolCallBlock(toolCall);
+                  const { block, state: state2 } = ensureToolCallBlock(toolCall);
                   if (!block.id && toolCall.id) {
                     block.id = toolCall.id;
                     toolCallBlocksById.set(toolCall.id, block);
@@ -41507,13 +41965,13 @@ var init_openai_completions = __esm({
                     block.name = name2;
                   }
                   let delta = "";
-                  if (toolCall.function?.arguments) {
+                  const input = state2.input;
+                  if (toolCall.function?.arguments && input.kind === "json") {
                     delta = toolCall.function.arguments;
-                    block.partialArgs = (block.partialArgs ?? "") + toolCall.function.arguments;
-                    block.arguments = parseStreamingJson(block.partialArgs);
-                  } else if (toolCall.custom?.input) {
-                    const nextInput = getCustomToolCallInput2(block) + toolCall.custom.input;
-                    delta = appendCustomToolCallInput2(block, nextInput, false) ?? "";
+                    block.arguments = input.arguments.append(delta);
+                  } else if (toolCall.custom?.input && input.kind === "custom") {
+                    delta = appendGrammarToolInputJsonDelta(input.buffer, input.property, toolCall.custom.input) ?? "";
+                    block.arguments = { [input.property]: input.buffer.input };
                   }
                   stream5.push({
                     type: "toolcall_delta",
@@ -41550,10 +42008,6 @@ var init_openai_completions = __esm({
             if (block.type === "thinking") {
               applyStreamedReasoningDetails(block);
             }
-            delete block.index;
-            delete block.partialArgs;
-            delete block.customInput;
-            delete block.streamIndex;
           }
           output.stopReason = options?.signal?.aborted ? "aborted" : "error";
           output.errorMessage = formatProviderError(normalizeProviderError(error));
@@ -41897,24 +42351,10 @@ function convertResponsesTools(tools, options) {
     return functionTool;
   });
 }
-function getCustomToolCallInput(block) {
-  const property2 = block.customInput?.property;
-  if (property2 === void 0)
-    return "";
-  const value = block.arguments[property2];
-  return typeof value === "string" ? value : "";
-}
-function appendCustomToolCallInput(block, nextInput, close) {
-  const customInput = block.customInput;
-  if (!customInput)
-    return void 0;
-  const delta = appendGrammarToolInputJsonDelta(customInput.jsonBuffer, customInput.property, nextInput, close);
-  block.arguments = { [customInput.property]: nextInput };
-  return delta;
-}
 async function processResponsesStream(openaiStream, output, stream5, model, options) {
   let sawTerminalResponseEvent = false;
   const outputSlots = /* @__PURE__ */ new Map();
+  const unfinishedToolCalls = /* @__PURE__ */ new Set();
   const reasoningBlocksById = /* @__PURE__ */ new Map();
   const applyMessagePhaseStopReason = (item) => {
     if (item.type === "message" && item.phase === "final_answer") {
@@ -41958,19 +42398,21 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
       return slot;
     }
     if (item.type === "function_call") {
+      const args = new ToolCallArgumentsStream();
       const block = {
         type: "toolCall",
         id: `${item.call_id}|${item.id}`,
         name: item.name,
-        arguments: {},
-        ...item.namespace !== void 0 ? { namespace: item.namespace } : {},
-        partialJson: item.arguments || ""
+        arguments: item.arguments ? args.append(item.arguments) : {},
+        ...item.namespace !== void 0 ? { namespace: item.namespace } : {}
       };
       output.content.push(block);
+      unfinishedToolCalls.add(block);
       const slot = {
         type: "toolCall",
         block,
-        contentIndex: output.content.length - 1
+        contentIndex: output.content.length - 1,
+        input: { kind: "json", arguments: args }
       };
       outputSlots.set(outputIndex, slot);
       stream5.push({ type: "toolcall_start", contentIndex: slot.contentIndex, partial: output });
@@ -41984,17 +42426,15 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
         id: `${item.call_id}|${item.id}`,
         name: item.name,
         arguments: { [inputProperty]: input },
-        ...item.namespace !== void 0 ? { namespace: item.namespace } : {},
-        customInput: {
-          property: inputProperty,
-          jsonBuffer: { input: "", started: false, closed: false }
-        }
+        ...item.namespace !== void 0 ? { namespace: item.namespace } : {}
       };
       output.content.push(block);
+      unfinishedToolCalls.add(block);
       const slot = {
         type: "toolCall",
         block,
-        contentIndex: output.content.length - 1
+        contentIndex: output.content.length - 1,
+        input: customInput(inputProperty, input)
       };
       outputSlots.set(outputIndex, slot);
       stream5.push({ type: "toolcall_start", contentIndex: slot.contentIndex, partial: output });
@@ -42124,33 +42564,28 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
       });
     } else if (event.type === "response.function_call_arguments.delta") {
       const slot = getSlot(event.output_index, "toolCall");
-      if (!slot || slot.block.partialJson === void 0)
+      if (slot?.input.kind !== "json")
         continue;
-      slot.block.partialJson += event.delta;
-      slot.block.arguments = parseStreamingJson(slot.block.partialJson);
+      slot.block.arguments = slot.input.arguments.append(event.delta);
       pushToolCallDelta(slot, event.delta);
     } else if (event.type === "response.function_call_arguments.done") {
       const slot = getSlot(event.output_index, "toolCall");
-      if (!slot || slot.block.partialJson === void 0)
+      if (slot?.input.kind !== "json")
         continue;
-      const previousPartialJson = slot.block.partialJson;
-      slot.block.partialJson = event.arguments;
-      slot.block.arguments = parseStreamingJson(slot.block.partialJson);
-      if (event.arguments.startsWith(previousPartialJson)) {
-        const delta = event.arguments.slice(previousPartialJson.length);
-        if (delta.length > 0)
-          pushToolCallDelta(slot, delta);
-      }
+      const replaced = slot.input.arguments.replace(event.arguments);
+      slot.block.arguments = replaced.arguments;
+      if (replaced.delta)
+        pushToolCallDelta(slot, replaced.delta);
     } else if (event.type === "response.custom_tool_call_input.delta") {
       const slot = getSlot(event.output_index, "toolCall");
-      if (!slot || !slot.block.customInput)
+      if (slot?.input.kind !== "custom")
         continue;
-      pushToolCallDelta(slot, appendCustomToolCallInput(slot.block, getCustomToolCallInput(slot.block) + event.delta, false));
+      pushToolCallDelta(slot, appendCustomInput(slot.block, slot.input, event.delta));
     } else if (event.type === "response.custom_tool_call_input.done") {
       const slot = getSlot(event.output_index, "toolCall");
-      if (!slot || !slot.block.customInput)
+      if (slot?.input.kind !== "custom")
         continue;
-      pushToolCallDelta(slot, appendCustomToolCallInput(slot.block, event.input, true));
+      pushToolCallDelta(slot, closeCustomInput(slot.block, slot.input, event.input));
     } else if (event.type === "response.output_item.done") {
       const item = event.item;
       applyMessagePhaseStopReason(item);
@@ -42178,11 +42613,13 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
           partial: output
         });
         outputSlots.delete(event.output_index);
-      } else if (item.type === "function_call" && slot?.type === "toolCall" && slot.block.partialJson !== void 0) {
-        Object.assign(slot.block, parseToolCallArguments(item.arguments || slot.block.partialJson));
+      } else if (item.type === "function_call" && slot?.type === "toolCall" && slot.input.kind === "json") {
+        if (item.arguments)
+          slot.input.arguments.replace(item.arguments);
+        Object.assign(slot.block, slot.input.arguments.finish());
         if (item.namespace !== void 0)
           slot.block.namespace = item.namespace;
-        delete slot.block.partialJson;
+        unfinishedToolCalls.delete(slot.block);
         stream5.push({
           type: "toolcall_end",
           contentIndex: slot.contentIndex,
@@ -42190,11 +42627,11 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
           partial: output
         });
         outputSlots.delete(event.output_index);
-      } else if (item.type === "custom_tool_call" && slot?.type === "toolCall" && slot.block.customInput) {
-        pushToolCallDelta(slot, appendCustomToolCallInput(slot.block, item.input ?? getCustomToolCallInput(slot.block), true));
+      } else if (item.type === "custom_tool_call" && slot?.type === "toolCall" && slot.input.kind === "custom") {
+        pushToolCallDelta(slot, closeCustomInput(slot.block, slot.input, item.input));
         if (item.namespace !== void 0)
           slot.block.namespace = item.namespace;
-        delete slot.block.customInput;
+        unfinishedToolCalls.delete(slot.block);
         stream5.push({
           type: "toolcall_end",
           contentIndex: slot.contentIndex,
@@ -42220,16 +42657,34 @@ async function processResponsesStream(openaiStream, output, stream5, model, opti
     throw new Error("OpenAI Responses stream ended before a terminal response event");
   }
   if (output.stopReason === "toolUse") {
-    for (const block of output.content) {
-      if (block.type !== "toolCall")
-        continue;
-      const toolCall = block;
-      if (toolCall.partialJson !== void 0 || toolCall.customInput !== void 0) {
-        throw new Error(`OpenAI Responses stream completed with an unfinished tool call: ${toolCall.name} (${toolCall.id})`);
-      }
+    for (const toolCall of unfinishedToolCalls) {
+      throw new Error(`OpenAI Responses stream completed with an unfinished tool call: ${toolCall.name} (${toolCall.id})`);
     }
   }
 }
+function customInput(property2, initialInput) {
+  const buffer = { input: "", started: false, closed: false };
+  return {
+    kind: "custom",
+    property: property2,
+    buffer,
+    unsent: appendGrammarToolInputJsonDelta(buffer, property2, initialInput) ?? ""
+  };
+}
+function appendCustomInput(block, input, delta) {
+  return sendCustomInput(block, input, appendGrammarToolInputJsonDelta(input.buffer, input.property, delta));
+}
+function closeCustomInput(block, input, finalInput) {
+  return sendCustomInput(block, input, closeGrammarToolInputJson(input.buffer, input.property, finalInput));
+}
+function sendCustomInput(block, input, jsonDelta) {
+  block.arguments = { [input.property]: input.buffer.input };
+  if (input.unsent === "")
+    return jsonDelta;
+  const sent = input.unsent + (jsonDelta ?? "");
+  input.unsent = "";
+  return sent;
+}
 function mapStopReason3(status, incompleteReason) {
   if (!status)
     return { stopReason: "stop" };
@@ -42261,8 +42716,8 @@ var init_openai_responses_shared = __esm({
   "node_modules/@earendil-works/pi-ai/dist/api/openai-responses-shared.js"() {
     init_models();
     init_hash();
-    init_json_parse();
     init_sanitize_unicode();
+    init_streaming_json();
     init_text();
     init_transcript();
     init_constrained_sampling();
@@ -42548,11 +43003,6 @@ var init_openai_responses = __esm({
           stream5.push({ type: "done", reason: output.stopReason, message: output });
           stream5.end();
         } catch (error) {
-          for (const block of output.content) {
-            delete block.index;
-            delete block.partialJson;
-            delete block.customInput;
-          }
           output.stopReason = options?.signal?.aborted ? "aborted" : "error";
           const errorMessage = formatProviderError(normalizeProviderError(error), `${model.provider === "openai" ? "OpenAI" : model.provider} API error`);
           if (errorMessage.includes("subscription_sharing_usage_limit_exceeded")) {
@@ -45404,7 +45854,7 @@ var require_webidl = __commonJS({
     var BOOLEAN = 2;
     var STRING = 3;
     var SYMBOL = 4;
-    var NUMBER = 5;
+    var NUMBER2 = 5;
     var BIGINT = 6;
     var NULL2 = 7;
     var OBJECT = 8;
@@ -45467,7 +45917,7 @@ var require_webidl = __commonJS({
         case "symbol":
           return SYMBOL;
         case "number":
-          return NUMBER;
+          return NUMBER2;
         case "bigint":
           return BIGINT;
         case "function":
@@ -45484,7 +45934,7 @@ var require_webidl = __commonJS({
       BOOLEAN,
       STRING,
       SYMBOL,
-      NUMBER,
+      NUMBER: NUMBER2,
       BIGINT,
       NULL: NULL2,
       OBJECT
@@ -45499,7 +45949,7 @@ var require_webidl = __commonJS({
           return "String";
         case SYMBOL:
           return "Symbol";
-        case NUMBER:
+        case NUMBER2:
           return "Number";
         case BIGINT:
           return "BigInt";
@@ -73692,7 +74142,7 @@ import { idleWatchdog, timeoutOf } from "@deepseek-ai/dsh-timeout";
 init_models();
 
 // node_modules/@earendil-works/pi-ai/dist/providers/data/opencode-go.json
-var opencode_go_default = { "anthropic-messages": { "chat:claude-haiku-5-5": { id: "claude-haiku-5-5", name: "Claude Haiku 5.5", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.5, cacheRead: 0.01, cacheWrite: 0.125, tiers: [{ inputTokensAbove: 1e5, input: 0.5, output: 2.5, cacheRead: 0.05, cacheWrite: 0.625 }] }, contextWindow: 1e6, maxTokens: 128e3, thinkingLevelMap: { xhigh: "xhigh", max: "max" }, compat: { forceAdaptiveThinking: true, supportsTemperature: false }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:minimax-m3": { id: "minimax-m3", name: "MiniMax-M3", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go", reasoning: true, input: ["text", "image"], cost: { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0, tiers: [{ inputTokensAbove: 512e3, input: 0.6, output: 2.4, cacheRead: 0.12, cacheWrite: 0 }] }, contextWindow: 1e6, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.7-plus": { id: "qwen3.7-plus", name: "Qwen3.7 Plus", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go", reasoning: true, input: ["text", "image"], cost: { input: 0.4, output: 1.6, cacheRead: 0.04, cacheWrite: 0.5, tiers: [{ inputTokensAbove: 256e3, input: 1.2, output: 4.8, cacheRead: 0.12, cacheWrite: 1.5 }] }, contextWindow: 1e6, maxTokens: 65536, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.8-flash": { id: "qwen3.8-flash", name: "Qwen3.8 Flash", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.47, cacheRead: 0.016, cacheWrite: 0.2 }, contextWindow: 1e6, maxTokens: 131072, compat: { allowEmptySignature: true }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.8-max": { id: "qwen3.8-max", name: "Qwen3.8 Max", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 6, cacheRead: 0.25, cacheWrite: 2.5 }, contextWindow: 1e6, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "google-generative-ai": { "chat:gemini-4-argon-contributor": { id: "gemini-4-argon-contributor", name: "Gemini 4 Argon Contributor", api: "google-generative-ai", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: null, max: null }, input: ["text", "image"], cost: { input: 4, output: 20, cacheRead: 0.2, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 65536, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "openai-completions": { "chat:beam-free": { id: "beam-free", name: "Beam Free", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1048576, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, type: "chat" }, "chat:deepseek-v4-pro": { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro (New)", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0.66, output: 1.98, cacheRead: 0.022, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", requiresReasoningContentOnAssistantMessages: true, thinkingFormat: "deepseek" }, contextWindow: 1e6, maxTokens: 384e3, thinkingLevelMap: { minimal: null, low: null, medium: null, high: "high", max: "max" }, type: "chat" }, "chat:deepseek-v4.1-flash": { id: "deepseek-v4.1-flash", name: "DeepSeek V4.1 Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, input: ["text", "image"], cost: { input: 0.15, output: 0.6, cacheRead: 3e-3, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", requiresReasoningContentOnAssistantMessages: true, thinkingFormat: "deepseek" }, contextWindow: 1e6, maxTokens: 384e3, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:glm-5.2": { id: "glm-5.2", name: "GLM-5.2", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: null, low: null, medium: null, high: "high", xhigh: null, max: "max" }, type: "chat" }, "chat:glm-5.3": { id: "glm-5.3", name: "GLM-5.3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, type: "chat" }, "chat:glm-5.3-flash": { id: "glm-5.3-flash", name: "GLM-5.3-Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.5, cacheRead: 0.03, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:hy3": { id: "hy3", name: "Hy3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0.14, output: 0.58, cacheRead: 0.035, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 256e3, maxTokens: 128e3, thinkingLevelMap: { off: "none", minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: null }, type: "chat" }, "chat:hy4-preview": { id: "hy4-preview", name: "Hy4 preview", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0.834, output: 2.501, cacheRead: 0.042, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1024e3, maxTokens: 64e3, thinkingLevelMap: { off: "none", minimal: null, low: null, medium: null, high: "high", xhigh: null, max: null }, type: "chat" }, "chat:kimi-k2.7-code": { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.95, output: 4, cacheRead: 0.19, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 262144, maxTokens: 262144, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:kimi-k3": { id: "kimi-k3", name: "Kimi K3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", supportsMidConvoSystemMessages: true, supportsMidConvoToolAdditions: true }, contextWindow: 1048576, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: null, low: null, medium: null, high: null, xhigh: null, max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:longcat-2.0": { id: "longcat-2.0", name: "LongCat-2.0", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0.3, output: 1.2, cacheRead: 6e-3, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 131072, type: "chat" }, "chat:longcat-2.5-preview-free": { id: "longcat-2.5-preview-free", name: "LongCat 2.5 Preview Free", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:mimo-v2.5": { id: "mimo-v2.5", name: "MiMo V2.5", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.14, output: 0.28, cacheRead: 28e-4, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 128e3, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:mimo-v2.5-pro": { id: "mimo-v2.5-pro", name: "MiMo V2.5 Pro", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text"], cost: { input: 0.435, output: 0.87, cacheRead: 3625e-6, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1048576, maxTokens: 128e3, type: "chat" }, "chat:mimo-v2.6-flash": { id: "mimo-v2.6-flash", name: "MiMo-V2.6-Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.14, output: 0.28, cacheRead: 28e-4, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1048576, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:mimo-v2.6-pro": { id: "mimo-v2.6-pro", name: "MiMo-V2.6-Pro", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.435, output: 0.87, cacheRead: 3625e-6, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1048576, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:minimax-m3.1-flash-free": { id: "minimax-m3.1-flash-free", name: "MiniMax M3.1 Flash Free", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 512e3, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:step-5-preview": { id: "step-5-preview", name: "Step 5 Preview", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 1, output: 2.7, cacheRead: 0.1, cacheWrite: 0 }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, contextWindow: 1e6, maxTokens: 65536, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: null, max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "openai-responses": { "chat:gpt-5.6-luna": { id: "gpt-5.6-luna", name: "GPT-5.6 Luna", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.2, output: 1.2, cacheRead: 0.02, cacheWrite: 0.25, tiers: [{ inputTokensAbove: 272e3, input: 0.4, output: 1.8, cacheRead: 0.04, cacheWrite: 0.5 }] }, compat: { sessionAffinityFormat: "openai-nosession", supportsMidConvoSystemMessages: true, supportsAdditionalTools: true }, contextWindow: 105e4, maxTokens: 128e3, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:gpt-6-luna": { id: "gpt-6-luna", name: "GPT-6 Luna", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.5, cacheRead: 0.01, cacheWrite: 0.125, tiers: [{ inputTokensAbove: 272e3, input: 0.2, output: 0.75, cacheRead: 0.02, cacheWrite: 0.25 }] }, compat: { sessionAffinityFormat: "openai-nosession", supportsMidConvoSystemMessages: true, supportsAdditionalTools: true }, contextWindow: 105e4, maxTokens: 128e3, thinkingLevelMap: { off: "none", minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:grok-4.7": { id: "grok-4.7", name: "Grok 4.7", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 6, cacheRead: 0.5, cacheWrite: 0, tiers: [{ inputTokensAbove: 2e5, input: 4, output: 12, cacheRead: 1, cacheWrite: 0 }] }, compat: { sessionAffinityFormat: "openai-nosession" }, contextWindow: 5e5, maxTokens: 5e5, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:muse-spark-1.2-contributor": { id: "muse-spark-1.2-contributor", name: "Muse Spark 1.2 Contributor", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.2, cacheRead: 2e-3, cacheWrite: 0 }, compat: { sessionAffinityFormat: "openai-nosession" }, contextWindow: 1048576, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: "minimal", low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:muse-spark-1.3-contributor": { id: "muse-spark-1.3-contributor", name: "Muse Spark 1.3 Contributor", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.2, cacheRead: 2e-3, cacheWrite: 0 }, compat: { sessionAffinityFormat: "openai-nosession" }, contextWindow: 1048576, maxTokens: 131072, thinkingLevelMap: { off: null, minimal: "minimal", low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } } };
+var opencode_go_default = { "anthropic-messages": { "chat:claude-haiku-5-5": { id: "claude-haiku-5-5", name: "Claude Haiku 5.5", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/anthropic", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.5, cacheRead: 0.01, cacheWrite: 0.125, tiers: [{ inputTokensAbove: 1e5, input: 0.5, output: 2.5, cacheRead: 0.05, cacheWrite: 0.625 }] }, contextWindow: 1e6, maxTokens: 128e3, thinkingLevelMap: { xhigh: "xhigh", max: "max" }, compat: { forceAdaptiveThinking: true, supportsTemperature: false }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:minimax-m3": { id: "minimax-m3", name: "MiniMax-M3", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/anthropic", reasoning: true, input: ["text", "image"], cost: { input: 0.3, output: 1.2, cacheRead: 0.06, cacheWrite: 0, tiers: [{ inputTokensAbove: 512e3, input: 0.6, output: 2.4, cacheRead: 0.12, cacheWrite: 0 }] }, contextWindow: 1e6, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.7-plus": { id: "qwen3.7-plus", name: "Qwen3.7 Plus", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/anthropic", reasoning: true, input: ["text", "image"], cost: { input: 0.4, output: 1.6, cacheRead: 0.04, cacheWrite: 0.5, tiers: [{ inputTokensAbove: 256e3, input: 1.2, output: 4.8, cacheRead: 0.12, cacheWrite: 1.5 }] }, contextWindow: 1e6, maxTokens: 65536, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.8-flash": { id: "qwen3.8-flash", name: "Qwen3.8 Flash", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/anthropic", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.47, cacheRead: 0.016, cacheWrite: 0.2 }, contextWindow: 1e6, maxTokens: 131072, compat: { allowEmptySignature: true }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:qwen3.8-max": { id: "qwen3.8-max", name: "Qwen3.8 Max", api: "anthropic-messages", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/anthropic", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 6, cacheRead: 0.25, cacheWrite: 2.5 }, contextWindow: 1e6, maxTokens: 131072, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "google-generative-ai": { "chat:gemini-4-argon": { id: "gemini-4-argon", name: "Gemini 4 Argon", api: "google-generative-ai", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/google/v1beta", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 10, cacheRead: 0.1, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 65536, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: null, max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "openai-completions": { "chat:beam-free": { id: "beam-free", name: "Beam Free", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, type: "chat" }, "chat:deepseek-v4-pro": { id: "deepseek-v4-pro", name: "DeepSeek V4 Pro (New)", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0.66, output: 1.98, cacheRead: 0.022, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 384e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", requiresReasoningContentOnAssistantMessages: true, thinkingFormat: "deepseek" }, thinkingLevelMap: { minimal: null, low: null, medium: null, high: "high", max: "max" }, type: "chat" }, "chat:deepseek-v4.1-flash": { id: "deepseek-v4.1-flash", name: "DeepSeek V4.1 Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.6, cacheRead: 3e-3, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 384e3, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", requiresReasoningContentOnAssistantMessages: true, thinkingFormat: "deepseek" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:glm-5.2": { id: "glm-5.2", name: "GLM-5.2", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: null, medium: null, high: "high", xhigh: null, max: "max" }, type: "chat" }, "chat:glm-5.3": { id: "glm-5.3", name: "GLM-5.3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, type: "chat" }, "chat:glm-5.3-flash": { id: "glm-5.3-flash", name: "GLM-5.3-Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.5, cacheRead: 0.03, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:hy3": { id: "hy3", name: "Hy3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0.14, output: 0.58, cacheRead: 0.035, cacheWrite: 0 }, contextWindow: 256e3, maxTokens: 128e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: "none", minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: null }, type: "chat" }, "chat:hy4-preview": { id: "hy4-preview", name: "Hy4 preview", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0.834, output: 2.501, cacheRead: 0.042, cacheWrite: 0 }, contextWindow: 1024e3, maxTokens: 64e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: "none", minimal: null, low: null, medium: null, high: "high", xhigh: null, max: null }, type: "chat" }, "chat:iceman-contributor-free": { id: "iceman-contributor-free", name: "Iceman Contributor Free", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 }, contextWindow: 262144, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, type: "chat" }, "chat:kimi-k2.7-code": { id: "kimi-k2.7-code", name: "Kimi K2.7 Code", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.95, output: 4, cacheRead: 0.19, cacheWrite: 0 }, contextWindow: 262144, maxTokens: 262144, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:kimi-k3": { id: "kimi-k3", name: "Kimi K3", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 3, output: 15, cacheRead: 0.3, cacheWrite: 3 }, contextWindow: 1048576, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", supportsMidConvoSystemMessages: true, supportsMidConvoToolAdditions: true }, thinkingLevelMap: { off: null, minimal: null, low: null, medium: null, high: null, xhigh: null, max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:kimi-k3.1-pro": { id: "kimi-k3.1-pro", name: "Kimi K3.1 Pro", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 10, cacheRead: 0.2, cacheWrite: 2 }, contextWindow: 1048576, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens", supportsMidConvoSystemMessages: true, supportsMidConvoToolAdditions: true }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: null, high: "high", xhigh: null, max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:longcat-2.0": { id: "longcat-2.0", name: "LongCat-2.0", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0.3, output: 1.2, cacheRead: 6e-3, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, type: "chat" }, "chat:mimo-v2.5": { id: "mimo-v2.5", name: "MiMo V2.5", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.14, output: 0.28, cacheRead: 28e-4, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 128e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:mimo-v2.5-pro": { id: "mimo-v2.5-pro", name: "MiMo V2.5 Pro", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text"], cost: { input: 0.435, output: 0.87, cacheRead: 3625e-6, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 128e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, type: "chat" }, "chat:mimo-v2.6-flash": { id: "mimo-v2.6-flash", name: "MiMo-V2.6-Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.14, output: 0.28, cacheRead: 28e-4, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:mimo-v2.6-pro": { id: "mimo-v2.6-pro", name: "MiMo-V2.6-Pro", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.435, output: 0.87, cacheRead: 3625e-6, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 131072, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:minimax-m3.1-flash": { id: "minimax-m3.1-flash", name: "MiniMax M3.1 Flash", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.15, output: 0.6, cacheRead: 0.015, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 512e3, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:step-5-preview": { id: "step-5-preview", name: "Step 5 Preview", api: "openai-completions", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 1, output: 2.7, cacheRead: 0.1, cacheWrite: 0 }, contextWindow: 1e6, maxTokens: 65536, compat: { supportsStore: false, supportsDeveloperRole: false, supportsStrictMode: true, maxTokensField: "max_tokens" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: null, max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } }, "openai-responses": { "chat:gpt-5.6-luna": { id: "gpt-5.6-luna", name: "GPT-5.6 Luna", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.2, output: 1.2, cacheRead: 0.02, cacheWrite: 0.25, tiers: [{ inputTokensAbove: 272e3, input: 0.4, output: 1.8, cacheRead: 0.04, cacheWrite: 0.5 }] }, contextWindow: 105e4, maxTokens: 128e3, compat: { sessionAffinityFormat: "openai-nosession", supportsMidConvoSystemMessages: true, supportsAdditionalTools: true }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:gpt-6-luna": { id: "gpt-6-luna", name: "GPT-6 Luna", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.5, cacheRead: 0.01, cacheWrite: 0.125, tiers: [{ inputTokensAbove: 272e3, input: 0.2, output: 0.75, cacheRead: 0.02, cacheWrite: 0.25 }] }, contextWindow: 105e4, maxTokens: 128e3, compat: { sessionAffinityFormat: "openai-nosession", supportsMidConvoSystemMessages: true, supportsAdditionalTools: true }, thinkingLevelMap: { off: "none", minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: "max" }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:grok-4.7": { id: "grok-4.7", name: "Grok 4.7", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 2, output: 6, cacheRead: 0.5, cacheWrite: 0, tiers: [{ inputTokensAbove: 2e5, input: 4, output: 12, cacheRead: 1, cacheWrite: 0 }] }, contextWindow: 5e5, maxTokens: 5e5, compat: { sessionAffinityFormat: "openai-nosession" }, thinkingLevelMap: { off: null, minimal: null, low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:muse-spark-1.2-contributor": { id: "muse-spark-1.2-contributor", name: "Muse Spark 1.2 Contributor", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.2, cacheRead: 2e-3, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 131072, compat: { sessionAffinityFormat: "openai-nosession" }, thinkingLevelMap: { off: null, minimal: "minimal", low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" }, "chat:muse-spark-1.3-contributor": { id: "muse-spark-1.3-contributor", name: "Muse Spark 1.3 Contributor", api: "openai-responses", provider: "opencode-go", baseUrl: "https://opencode.ai/inference/go/openai/v1", reasoning: true, input: ["text", "image"], cost: { input: 0.1, output: 0.2, cacheRead: 2e-3, cacheWrite: 0 }, contextWindow: 1048576, maxTokens: 131072, compat: { sessionAffinityFormat: "openai-nosession" }, thinkingLevelMap: { off: null, minimal: "minimal", low: "low", medium: "medium", high: "high", xhigh: "xhigh", max: null }, inputLimits: { images: { resize: { maxWidth: 2e3, maxHeight: 2e3, maxBytes: 4718592, jpegQuality: 80 } } }, type: "chat" } } };
 
 // node_modules/@earendil-works/pi-ai/dist/model-catalog.js
 function flattenModelCatalog(groups, type) {
@@ -77574,6 +78024,8 @@ var OVERFLOW_PATTERNS = [
   // OpenAI (Completions & Responses API)
   /exceeds (?:the )?(?:model'?s )?maximum context length(?: of [\d,]+ tokens?|\s*\([\d,]+\))/i,
   // OpenAI-compatible proxies (LiteLLM)
+  /context window exceeded/i,
+  // LiteLLM pre-call check (ContextWindowExceededError)
   /input token count.*exceeds the maximum/i,
   // Google (Gemini)
   /maximum prompt length is \d+/i,
@@ -79646,7 +80098,7 @@ function createUpdateChecker(deps) {
 }
 
 // src/update/version.ts
-var PLUGIN_VERSION = true ? "0.1.88" : "0.0.0-dev";
+var PLUGIN_VERSION = true ? "0.1.89" : "0.0.0-dev";
 
 // src/update/service.ts
 import { TypertRemoteService as TypertRemoteService4 } from "@deepseek-ai/dsh-typert-protocol";
@@ -79968,9 +80420,11 @@ function apply(ctx, raw) {
     undiscover();
     return closeDispatcher();
   });
-  ctx.on("internal/config", function(_raw, next) {
-    const value = next();
-    if (this === ctx.fiber) assertBaseURL(PlainConfig(value).baseURL);
+  ctx.on("internal/config", function(first, second) {
+    const value = second();
+    if (ctx.fiber !== void 0 && (first === ctx.fiber || this === ctx.fiber)) {
+      assertBaseURL(PlainConfig(value).baseURL);
+    }
     return value;
   });
   ctx.on("loader/volatile-update", syncRoute);