Merge commit '2fdcbb423bb60a63c973e9ed91ec3d095d8ff05a' into minh/realtime_prompting_guide

Author: minh-hoque

articles/gpt-oss/fine-tune-korean.ipynb

@@ -6,7 +6,7 @@ The [`gpt-oss` models](https://openai.com/open-models) were trained on the harmo
 
 ### Roles
 
-Every message that the model processes has a role associated with it. The model knows about three types of roles:
+Every message that the model processes has a role associated with it. The model knows about five types of roles:
 
 | Role        | Purpose                                                                                                                                                                                 |
 | :---------- | :-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
@@ -52,24 +52,15 @@ encoding = load_harmony_encoding(HarmonyEncodingName.HARMONY_GPT_OSS)
 
 system_message = (
     SystemContent.new()
-        .with_model_identity(
-            "You are ChatGPT, a large language model trained by OpenAI."
-        )
         .with_reasoning_effort(ReasoningEffort.HIGH)
         .with_conversation_start_date("2025-06-28")
-        .with_knowledge_cutoff("2024-06")
-        .with_required_channels(["analysis", "commentary", "final"])
 )
 
 developer_message = (
     DeveloperContent.new()
         .with_instructions("Always respond in riddles")
         .with_function_tools(
             [
-                ToolDescription.new(
-                    "get_location",
-                    "Gets the location of the user.",
-                ),
                 ToolDescription.new(
                     "get_current_weather",
                     "Gets the current weather in the provided location.",
@@ -90,7 +81,7 @@ developer_message = (
                     },
                 ),
             ]
-        )
+	)
 )
 
 convo = Conversation.from_messages(
@@ -100,16 +91,16 @@ convo = Conversation.from_messages(
         Message.from_role_and_content(Role.USER, "What is the weather in Tokyo?"),
         Message.from_role_and_content(
             Role.ASSISTANT,
-            'User asks: "What is the weather in Tokyo?" We need to use get_weather tool.',
+            'User asks: "What is the weather in Tokyo?" We need to use get_current_weather tool.',
         ).with_channel("analysis"),
         Message.from_role_and_content(Role.ASSISTANT, '{"location": "Tokyo"}')
         .with_channel("commentary")
-        .with_recipient("functions.get_weather")
-        .with_content_type("json"),
+        .with_recipient("functions.get_current_weather")
+        .with_content_type("<|constrain|> json"),
         Message.from_author_and_content(
-            Author.new(Role.TOOL, "functions.lookup_weather"),
+            Author.new(Role.TOOL, "functions.get_current_weather"),
             '{ "temperature": 20, "sunny": true }',
-        ).with_recipient("assistant").with_channel("commentary"),
+        ).with_channel("commentary"),
     ]
 )
 
@@ -202,6 +193,8 @@ Once its done generating it will stop with either a `<|return|>` token indicatin
 
 The `final` channel will contain the answer to your user’s request. Check out the [reasoning section](#reasoning) for more details on the chain-of-thought.
 
+**Implementation note:** `<|return|>` is a decode-time stop token only. When you add the assistant’s generated reply to conversation history for the next turn, replace the trailing `<|return|>` with `<|end|>` so that stored messages are fully formed as `<|start|>{header}<|message|>{content}<|end|>`. Prior messages in prompts should therefore end with `<|end|>`. For supervised targets/training examples, ending with `<|return|>` is appropriate; for persisted history, normalize to `<|end|>`.
+
 ### System message format
 
 The system message is used to provide general information to the system. This is different to what might be considered the “system prompt” in other prompt formats. For that, check out the [developer message format](#developer-message-format).
@@ -383,7 +376,7 @@ If the model decides to call a tool it will define a `recipient` in the header o
 The model might also specify a `<|constrain|>` token to indicate the type of input for the tool call. In this case since it’s being passed in as JSON the `<|constrain|>` is set to `json`.
 
 ```
-<|channel|>analysis<|message|>Need to use function get_weather.<|end|><|start|>assistant<|channel|>commentary to=functions.get_weather <|constrain|>json<|message|>{"location":"San Francisco"}<|call|>
+<|channel|>analysis<|message|>Need to use function get_current_weather.<|end|><|start|>assistant<|channel|>commentary to=functions.get_current_weather <|constrain|>json<|message|>{"location":"San Francisco"}<|call|>
 ```
 
 #### Handling tool calls
@@ -399,7 +392,7 @@ A tool message has the following format:
 So in our example above
 
 ```
-<|start|>functions.get_weather to=assistant<|channel|>commentary<|message|>{"sunny": true, "temperature": 20}<|end|>
+<|start|>functions.get_current_weather to=assistant<|channel|>commentary<|message|>{"sunny": true, "temperature": 20}<|end|>
 ```
 
 Once you have gathered the output for the tool calls you can run inference with the complete content:
@@ -439,10 +432,10 @@ locations: string[],
 format?: "celsius" | "fahrenheit", // default: celsius
 }) => any;
 
-} // namespace functions<|end|><|start|>user<|message|>What is the weather like in SF?<|end|><|start|>assistant<|channel|>analysis<|message|>Need to use function get_weather.<|end|><|start|>assistant<|channel|>commentary to=functions.get_weather <|constrain|>json<|message|>{"location":"San Francisco"}<|call|><|start|>functions.get_weather to=assistant<|channel|>commentary<|message|>{"sunny": true, "temperature": 20}<|end|><|start|>assistant
+} // namespace functions<|end|><|start|>user<|message|>What is the weather like in SF?<|end|><|start|>assistant<|channel|>analysis<|message|>Need to use function get_current_weather.<|end|><|start|>assistant<|channel|>commentary to=functions.get_current_weather <|constrain|>json<|message|>{"location":"San Francisco"}<|call|><|start|>functions.get_current_weather to=assistant<|channel|>commentary<|message|>{"sunny": true, "temperature": 20}<|end|><|start|>assistant
 ```
 
-As you can see above we are passing not just the function out back into the model for further sampling but also the previous chain-of-thought (“Need to use function get_weather.”) to provide the model with the necessary information to continue its chain-of-thought or provide the final answer.
+As you can see above we are passing not just the function out back into the model for further sampling but also the previous chain-of-thought (“Need to use function get_current_weather.”) to provide the model with the necessary information to continue its chain-of-thought or provide the final answer.
 
 #### Preambles
 

articles/run-nvidia.ipynb renamed to articles/gpt-oss/run-nvidia.ipynb

@@ -461,3 +461,8 @@ hendrytl:
   name: "Todd Hendry"
   website: "https://www.linkedin.com/in/todd-hendry-962aa577/"
   avatar: "https://avatars.githubusercontent.com/u/36863669"
+
+heejingithub:
+  name: "Heejin Cho"
+  website: "https://www.linkedin.com/in/heejc/"
+  avatar: "https://avatars.githubusercontent.com/u/169293861"

articles/openai-harmony.md

@@ -51,7 +51,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
+   "execution_count": null,
    "id": "dab872c5",
    "metadata": {
     "ExecuteTime": {
@@ -66,7 +66,7 @@
     "from tenacity import retry, wait_random_exponential, stop_after_attempt\n",
     "from termcolor import colored  \n",
     "\n",
-    "GPT_MODEL = \"gpt-4o\"\n",
+    "GPT_MODEL = \"gpt-5\"\n",
     "client = OpenAI()"
    ]
   },
@@ -515,7 +515,7 @@
    "source": [
     "### Parallel Function Calling\n",
     "\n",
-    "Newer models such as gpt-4o or gpt-3.5-turbo can call multiple functions in one turn."
+    "Newer models such as gpt-5, gpt-4.1 or gpt-4o can call multiple functions in one turn."
    ]
   },
   {
@@ -758,7 +758,7 @@
    "source": [
     "##### Steps to invoke a function call using Chat Completions API: \n",
     "\n",
-    "**Step 1**: Prompt the model with content that may result in model selecting a tool to use. The description of the tools such as a function names and signature is defined in the 'Tools' list and passed to the model in API call. If selected, the function name and parameters are included in the response.<br>\n",
+    "**Step 1**: Prompt the model with content that may result in model selecting a tool to use. The description of the tools such as a function name and signature is defined in the 'Tools' list and passed to the model in API call. If selected, the function name and parameters are included in the response.<br>\n",
     "  \n",
     "**Step 2**: Check programmatically if model wanted to call a function. If true, proceed to step 3. <br>  \n",
     "**Step 3**: Extract the function name and parameters from response, call the function with parameters. Append the result to messages. <br>    \n",
@@ -767,7 +767,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
+   "execution_count": null,
    "id": "e8b7cb9cdc7a7616",
    "metadata": {
     "ExecuteTime": {
@@ -792,9 +792,9 @@
     "}]\n",
     "\n",
     "response = client.chat.completions.create(\n",
-    "    model='gpt-4o', \n",
+    "    model=GPT_MODEL, \n",
     "    messages=messages, \n",
-    "    tools= tools, \n",
+    "    tools=tools, \n",
     "    tool_choice=\"auto\"\n",
     ")\n",
     "\n",
@@ -807,7 +807,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
+   "execution_count": null,
    "id": "351c39def3417776",
    "metadata": {
     "ExecuteTime": {
@@ -847,7 +847,7 @@
     "        # Step 4: Invoke the chat completions API with the function response appended to the messages list\n",
     "        # Note that messages with role 'tool' must be a response to a preceding message with 'tool_calls'\n",
     "        model_response_with_function_call = client.chat.completions.create(\n",
-    "            model=\"gpt-4o\",\n",
+    "            model=GPT_MODEL,\n",
     "            messages=messages,\n",
     "        )  # get a new response from the model where it can see the function response\n",
     "        print(model_response_with_function_call.choices[0].message.content)\n",

authors.yaml

@@ -253,7 +253,7 @@ async def process_api_requests_from_file(
                     )
                     await asyncio.sleep(remaining_seconds_to_pause)
                     # ^e.g., if pause is 15 seconds and final limit was hit 5 seconds ago
-                    logging.warn(
+                    logging.warning(
                         f"Pausing to cool down until {time.ctime(status_tracker.time_of_last_rate_limit_error + seconds_to_pause_after_rate_limit_error)}"
                     )
 

examples/How_to_call_functions_with_chat_models.ipynb

@@ -190,7 +190,7 @@
     "You can clone the openai-cookbook repository & take the sample python code & SAM template from the `lambda-middleware` directory:\n",
     "\n",
     "```\n",
-    "git clone https://github.com/pap-openai/lambda-middleware\n",
+    "git clone https://github.com/pap-openai/aws-lambda-middleware\n",
     "cd lambda-middleware\n",
     "```\n",
     "\n",

examples/api_request_parallel_processor.py

@@ -14,7 +14,7 @@
         "\n",
         "This cookbook provides a practical, step-by-step approach to automating the workflow between Jira and GitHub. By labeling a Jira issue, you trigger an end-to-end process that creates a **GitHub pull request**, keeps both systems updated, and streamlines code review, all with minimal manual effort. The automation is powered by the [`codex-cli`](https://github.com/openai/openai-codex) agent running inside a GitHub Action.\n",
         "\n",
-        "# <img src=\"../../images/codex_action.png\" alt=\"Full data-flow diagram\" width=\"500\"/>\n"
+        "<img src=\"../../images/codex_action.png\" alt=\"Full data-flow diagram\" width=\"500\"/>\n"
       ]
     },
     {