feat(openai): Properly pass through max_completion_tokens (#7683)

libs/langchain-openai/src/chat_models.ts

@@ -153,7 +153,7 @@ export function _convertMessagesToOpenAIParams(
   // TODO: Function messages do not support array content, fix cast
   return messages.flatMap((message) => {
     let role = messageToOpenAIRole(message);
-    if (role === "system" && model?.startsWith("o1")) {
+    if (role === "system" && isReasoningModel(model)) {
       role = "developer";
     }
     // eslint-disable-next-line @typescript-eslint/no-explicit-any
@@ -224,6 +224,10 @@ function _convertChatOpenAIToolTypeToOpenAITool(
   return _convertToOpenAITool(tool, fields);
 }
 
+function isReasoningModel(model?: string) {
+  return model?.startsWith("o1") || model?.startsWith("o3");
+}
+
 // TODO: Use the base structured output options param in next breaking release.
 export interface ChatOpenAIStructuredOutputMethodOptions<
   IncludeRaw extends boolean
@@ -1027,7 +1031,6 @@ export class ChatOpenAI<
     this.topP = fields?.topP ?? this.topP;
     this.frequencyPenalty = fields?.frequencyPenalty ?? this.frequencyPenalty;
     this.presencePenalty = fields?.presencePenalty ?? this.presencePenalty;
-    this.maxTokens = fields?.maxTokens;
     this.logprobs = fields?.logprobs;
     this.topLogprobs = fields?.topLogprobs;
     this.n = fields?.n ?? this.n;
@@ -1039,6 +1042,7 @@ export class ChatOpenAI<
     this.audio = fields?.audio;
     this.modalities = fields?.modalities;
     this.reasoningEffort = fields?.reasoningEffort;
+    this.maxTokens = fields?.maxCompletionTokens ?? fields?.maxTokens;
 
     if (this.model === "o1") {
       this.disableStreaming = true;
@@ -1151,7 +1155,6 @@ export class ChatOpenAI<
       top_p: this.topP,
       frequency_penalty: this.frequencyPenalty,
       presence_penalty: this.presencePenalty,
-      max_tokens: this.maxTokens === -1 ? undefined : this.maxTokens,
       logprobs: this.logprobs,
       top_logprobs: this.topLogprobs,
       n: this.n,
@@ -1187,6 +1190,12 @@ export class ChatOpenAI<
     if (reasoningEffort !== undefined) {
       params.reasoning_effort = reasoningEffort;
     }
+    if (isReasoningModel(params.model)) {
+      params.max_completion_tokens =
+        this.maxTokens === -1 ? undefined : this.maxTokens;
+    } else {
+      params.max_tokens = this.maxTokens === -1 ? undefined : this.maxTokens;
+    }
     return params;
   }
 

libs/langchain-openai/src/tests/chat_models.int.test.ts

@@ -1233,6 +1233,26 @@ test("Allows developer messages with o1", async () => {
   expect(res.content).toEqual("testing");
 });
 
+test("Works with maxCompletionTokens with o3", async () => {
+  const model = new ChatOpenAI({
+    model: "o3-mini",
+    reasoningEffort: "low",
+    maxCompletionTokens: 100,
+  });
+  const res = await model.invoke([
+    {
+      role: "system",
+      content: `Always respond only with the word "testing"`,
+    },
+    {
+      role: "user",
+      content: "hi",
+    },
+  ]);
+  console.log(res);
+  expect(res.content).toEqual("testing");
+});
+
 test.skip("Allow overriding", async () => {
   class ChatDeepSeek extends ChatOpenAI {
     protected override _convertOpenAIDeltaToBaseMessageChunk(

libs/langchain-openai/src/types.ts

@@ -23,6 +23,13 @@ export declare interface OpenAIBaseInput {
    */
   maxTokens?: number;
 
+  /**
+   * Maximum number of tokens to generate in the completion. -1 returns as many
+   * tokens as possible given the prompt and the model's maximum context size.
+   * Alias for `maxTokens` for reasoning models.
+   */
+  maxCompletionTokens?: number;
+
   /** Total probability mass of tokens to consider at each step */
   topP: number;