Struct OpenAIChatModel 

pub struct OpenAIChatModel { /* private fields */ }

OpenAI chat completion implementation of the language model interface.

This struct provides support for OpenAI’s chat completion API, enabling text generation, streaming, tool calling, and multi-modal capabilities. It handles conversion between the unified SDK format and OpenAI’s API specifications, including model-specific behaviors for reasoning models, search preview models, and flex processing.

Supported Features

• Text generation and streaming responses
• Function calling and tool use
• Multi-modal inputs (images, audio)
• JSON schema-based structured outputs
• Vision capabilities for supported models
• Reasoning effort customization for o1/o3 models
• Logit bias and probability tracking
• Token usage tracking including cached and reasoning tokens

OpenAI-Specific Behaviors

• System messages are converted to developer messages for reasoning models (o1, o3)
• Temperature configuration is ignored for search preview models with a warning
• Reasoning models use max_completion_tokens instead of max_tokens
• Flex processing service tier validation for supported models
• Streaming includes usage information in the final chunk

Implementations

impl OpenAIChatModel

pub fn new(model_id: impl Into<String>, config: impl Into<OpenAIConfig>) -> Self

Creates a new OpenAI chat model instance.

Initializes a language model that communicates with OpenAI’s chat completion API. The model will automatically detect capabilities and apply model-specific configurations based on the provided model ID (e.g., reasoning for o1/o3, special handling for vision models).

Arguments

• model_id - The OpenAI model identifier (e.g., “gpt-4o”, “gpt-4-turbo”, “gpt-4”, “gpt-3.5-turbo”, “o1”, “o1-mini”, “o3-mini”)
• config - OpenAI configuration containing API endpoint, authentication headers, and optional customizations

OpenAI Model Categories

• GPT-4o Series: Latest high-capability models with vision and reasoning
• Reasoning Models (o1/o3): Enhanced reasoning with special parameter handling
• GPT-4 Series: Advanced reasoning and code generation
• GPT-3.5 Turbo: Balanced performance and cost-effectiveness
• Search Models: Integration with web search capabilities

Example

use ai_sdk_openai::{OpenAIConfig, OpenAIChatModel};

let api_key = std::env::var("OPENAI_API_KEY").unwrap();
let config = OpenAIConfig::from_api_key(api_key);
let model = OpenAIChatModel::new("gpt-4o", config);

Trait Implementations

impl LanguageModel for OpenAIChatModel

fn provider(&self) -> &str

Returns the provider identifier for this model.

fn model_id(&self) -> &str

Returns the specific model identifier.

fn supported_urls<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = HashMap<String, Vec<String>>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Returns URLs supported by this model for various operations.

fn do_generate<'life0, 'async_trait>( &'life0 self, options: CallOptions, ) -> Pin<Box<dyn Future<Output = Result<GenerateResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Executes a non-streaming generation request.

fn do_stream<'life0, 'async_trait>( &'life0 self, options: CallOptions, ) -> Pin<Box<dyn Future<Output = Result<StreamResponse>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Executes a streaming generation request.

fn specification_version(&self) -> &str

Returns the specification version implemented by this model.

fn generate<P>(&self, prompt: P) -> GenerateBuilder<'_, Self>

where Self: Sized, P: TryInto<Prompt>, Box<dyn Error + Sync + Send>: From<<P as TryInto<Prompt>>::Error>,

Creates a builder for a non-streaming generation request.

fn stream<P>(&self, prompt: P) -> StreamBuilder<'_, Self>

where Self: Sized, P: TryInto<Prompt>, Box<dyn Error + Sync + Send>: From<<P as TryInto<Prompt>>::Error>,

Creates a builder for a streaming generation request.