# llm_openai A RocketRide LLM node that connects OpenAI GPT models to a pipeline. ## What it does Provides an OpenAI-backed chat LLM (`classType: llm`, capability `invoke`). It is used primarily as an `llm` invoke connection by agents and other nodes that need a language model, and can also be wired directly via lanes: send a question in, receive a generated answer out. Built on **langchain-openai** (`ChatOpenAI`) with the **openai** SDK underneath. Non-reasoning models are invoked through the Chat Completions API with `temperature: 0` and `max_tokens` set to the profile's output-token limit, so responses are deterministic by default. Reasoning-capable models (flagged via `capabilities.reasoning` in the model configuration) are routed through the **OpenAI Responses API** instead, using `max_completion_tokens`. This lets the node stream the model's reasoning summary over the `thinking` SSE lane in addition to the answer (`SUPPORTS_REASONING_STREAMING` is on). When a pipeline is validated, the node makes a tiny live test call to the configured model (a one-word prompt capped at a few tokens) to verify the API key and model name up front. Models whose names start with `gpt-5` (including `gpt-5.1`, `gpt-5-mini`, `gpt-5-nano`, and later families) are probed with `max_completion_tokens`; older ones with `max_tokens`. Validation failures surface as warnings with the provider's status code, error type, and message. Rate-limit and connection errors are treated as retryable; authentication and other API errors are not, and are mapped to friendly messages (e.g. `Invalid API key.`). --- ## Configuration ### Lanes | Lane in | Lane out | Description | | ----------- | --------- | ---------------------------------------------------- | | `questions` | `answers` | Send a question directly, receive a generated answer | ### Fields | Field | Type | Description | | ------------------ | ------ | ---------------------------------------------------------------------------------------- | | `profile` | enum | Model selection. Default `openai-5-2` (GPT-5.2). See profiles below. | | `apikey` | string | OpenAI API key. Required for every profile. | | `model` | string | OpenAI model name. Only editable in the `custom` profile. | | `modelTotalTokens` | number | Total token (context) limit. Only editable in the `custom` profile (default `16384`). Must be greater than 0. | | `modelSource` | enum | Where the model is hosted (standard cloud-LLM field, default `provider`). | Preconfigured profiles only expose `apikey` and `modelSource`; the model name and token limits come from the profile. The `custom` profile additionally exposes `model` and `modelTotalTokens` for any OpenAI model not in the list. --- ## Profiles Default profile: **`openai-5-2` (GPT-5.2)**. | Profile | Model | Context tokens | Output tokens | | ------------------------ | --------------------- | -------------- | ------------- | | `openai-5-4` | `gpt-5.4` | 1,050,000 | 128,000 | | `openai-5-4-pro` | `gpt-5.4-pro` | 1,050,000 | 128,000 | | `openai-5-4-mini` | `gpt-5.4-mini` | 400,000 | 128,000 | | `openai-5-4-nano` | `gpt-5.4-nano` | 400,000 | 128,000 | | `gpt-5-5` | `gpt-5.5` | 1,050,000 | 128,000 | | `openai-5-2` _(default)_ | `gpt-5.2` | 400,000 | 128,000 | | `openai-5-1` | `gpt-5.1` | 400,000 | 128,000 | | `openai-5` | `gpt-5` | 400,000 | 128,000 | | `openai-5-mini` | `gpt-5-mini` | 400,000 | 128,000 | | `openai-5-nano` | `gpt-5-nano` | 400,000 | 128,000 | | `gpt-5-3-chat-latest` | `gpt-5.3-chat-latest` | 16,384 | 16,384 | | `gpt-5-2-chat-latest` | `gpt-5.2-chat-latest` | 16,384 | 16,384 | | `gpt-5-1-chat-latest` | `gpt-5.1-chat-latest` | 16,384 | 16,384 | | `gpt-5-chat-latest` | `gpt-5-chat-latest` | 16,384 | 16,384 | | `openai-4o` | `gpt-4o` | 128,000 | 16,384 | | `openai-4o-mini` | `gpt-4o-mini` | 128,000 | 16,384 | | `gpt-4-1` | `gpt-4.1` | 1,047,576 | 32,768 | | `gpt-4-1-mini` | `gpt-4.1-mini` | 1,047,576 | 32,768 | | `gpt-4-1-nano` | `gpt-4.1-nano` | 1,047,576 | 32,768 | | `gpt-4-turbo` | `gpt-4-turbo` | 128,000 | 4,096 | | `gpt-4` | `gpt-4` | 8,191 | 4,096 | | `gpt-3-5-turbo` | `gpt-3.5-turbo` | 16,385 | 4,096 | | `gpt-3-5-turbo-16k` | `gpt-3.5-turbo-16k` | 16,385 | 4,096 | | `o1` | `o1` | 200,000 | 100,000 | | `o3` | `o3` | 200,000 | 100,000 | | `o3-mini` | `o3-mini` | 200,000 | 100,000 | | `o4-mini` | `o4-mini` | 200,000 | 100,000 | | `custom` | _(user-specified)_ | configurable (default 16,384) | n/a | --- ## Reasoning streaming For models whose configuration carries `capabilities.reasoning`, the node builds a raw `openai` client alongside the LangChain one and streams answers via the Responses API with `reasoning: {summary: 'auto'}`. Reasoning-summary deltas are forwarded over the `thinking` SSE lane while the answer streams normally. Non-reasoning models use plain LangChain streaming and are unaffected. --- ## Authentication Set `apikey` to an OpenAI API key for every profile (including `custom`). The key is verified with a live one-word test call during pipeline validation, so a bad key or model name is reported before the pipeline runs rather than at first invoke. --- ## Testing Automated node tests are declared in `services.json`: - **Mock group**: runs the `openai-5-2`, `openai-5-4`, `openai-5-4-pro`, `openai-5-4-mini`, and `openai-5-4-nano` profiles against a mocked `langchain_openai`; no real API key needed. `ROCKETRIDE_MOCK` must point to `nodes/test/mocks`. - **Real group**: skipped unless `ROCKETRIDE_OPENAI_KEY` is set; calls the live API with the `openai-4o-mini` profile. --- ## Schema | Field | Type | Description | Default | |---|---|---|---| | `model` | `string` | **Model**
OpenAI model | | | `modelTotalTokens` | `number` | **Tokens**
Total Tokens | | | `openai.profile` | `string` | **Model**
LLM model | `"openai-5-2"` | ## Dependencies - `openai` - `langchain-openai` - `langchain-core` - `langchain` ## Source [ View source](https://github.com/rocketride-org/rocketride-server/tree/develop/nodes/src/nodes/llm_openai)