Struct ChatCompletionRequest 

pub struct ChatCompletionRequest {

    pub model: String,
    pub messages: Vec<Message>,
    pub temperature: Option<f64>,
    pub top_p: Option<f64>,
    pub top_k: Option<u32>,
    pub max_tokens: Option<u32>,
    pub stream: Option<bool>,
    pub stop: Option<Value>,
    pub seed: Option<i64>,
    pub frequency_penalty: Option<f64>,
    pub presence_penalty: Option<f64>,
    pub repetition_penalty: Option<f64>,
    pub logit_bias: Option<Value>,
    pub logprobs: Option<bool>,
    pub top_logprobs: Option<u32>,
    pub min_p: Option<f64>,
    pub top_a: Option<f64>,
    pub tools: Option<Vec<Tool>>,
    pub tool_choice: Option<ToolChoice>,
    pub response_format: Option<ResponseFormat>,
    pub provider: Option<Provider>,
    pub reasoning: Option<ReasoningConfig>,
    pub transforms: Option<Vec<String>>,
    pub plugins: Option<Vec<Plugin>>,
    pub usage: Option<Value>,
    pub user: Option<String>,
}

Chat-completions request payload.

Fields

model: String

Model id (e.g. google/gemini-3.1-flash-lite).

messages: Vec<Message>

Conversation messages.

temperature: Option<f64>

Sampling temperature (0.0 = deterministic, ~1.0 = creative).

top_p: Option<f64>

Nucleus sampling probability.

top_k: Option<u32>

Cap on tokens considered each step.

max_tokens: Option<u32>

Max generated tokens.

stream: Option<bool>

Stream-mode flag. The SDK forces Some(true) for streaming methods and Some(false) for blocking methods.

stop: Option<Value>

Stop sequence(s) — string or array of strings.

seed: Option<i64>

Sampling seed for reproducibility.

frequency_penalty: Option<f64>

Frequency penalty.

presence_penalty: Option<f64>

Presence penalty.

repetition_penalty: Option<f64>

Repetition penalty.

logit_bias: Option<Value>

Per-token bias map (provider-specific shape).

logprobs: Option<bool>

Request log-probabilities for the chosen tokens.

top_logprobs: Option<u32>

Number of alternate tokens to report logprobs for.

min_p: Option<f64>

Min-p sampling threshold.

top_a: Option<f64>

Top-a sampling threshold.

tools: Option<Vec<Tool>>

Tools the model may call.

tool_choice: Option<ToolChoice>

Tool-selection strategy.

response_format: Option<ResponseFormat>

Output format constraint (JSON schema / JSON mode).

provider: Option<Provider>

Provider-routing parameters.

reasoning: Option<ReasoningConfig>

Reasoning-token configuration.

transforms: Option<Vec<String>>

Message-transform pipeline (e.g. ["middle-out"]).

plugins: Option<Vec<Plugin>>

Plugin configuration (web search, file parsing, …).

usage: Option<Value>

Provider-specific usage reporting hint.

user: Option<String>

Opaque end-user identifier for abuse monitoring.

Implementations

impl ChatCompletionRequest

pub fn new(model: impl Into<String>, messages: Vec<Message>) -> Self

Construct a new request with just the required fields.

pub fn with_tools(self, tools: Vec<Tool>) -> Self

Attach the list of tools the model may call.

pub fn with_tool_choice(self, choice: ToolChoice) -> Self

Set the tool-selection strategy.

pub fn with_response_format(self, format: ResponseFormat) -> Self

Set the response format directly.

pub fn with_json_schema( self, name: impl Into<String>, strict: bool, schema: Value, ) -> Self

Constrain responses to a named JSON schema. Sugar over with_response_format(ResponseFormat::json_schema(...)).

pub fn with_json_mode(self) -> Self

Ask the model to emit a JSON object (no schema). Sugar over with_response_format(ResponseFormat::json_object()).

pub fn with_transforms<S, I>(self, transforms: I) -> Self

where S: Into<String>, I: IntoIterator<Item = S>,

Set the message-transform pipeline. Passing an empty slice explicitly disables OpenRouter’s default transforms.

pub fn with_provider(self, provider: Provider) -> Self

Replace the provider-routing config with provider.

pub fn with_provider_order<S, I>(self, order: I) -> Self

where S: Into<String>, I: IntoIterator<Item = S>,

Preferred provider order.

pub fn with_provider_sort(self, sort: impl Into<String>) -> Self

Sort strategy: "throughput", "price", or "latency".

pub fn with_only_providers<S, I>(self, only: I) -> Self

where S: Into<String>, I: IntoIterator<Item = S>,

Restrict to this set of providers.

pub fn with_ignore_providers<S, I>(self, ignore: I) -> Self

where S: Into<String>, I: IntoIterator<Item = S>,

Exclude these providers.

pub fn with_quantizations<S, I>(self, q: I) -> Self

where S: Into<String>, I: IntoIterator<Item = S>,

Permitted quantization tiers.

pub fn with_max_price(self, price: Value) -> Self

Per-token max-price filter.

pub fn with_data_collection(self, policy: impl Into<String>) -> Self

Data-collection policy: "allow" or "deny".

pub fn with_require_parameters(self, required: bool) -> Self

Require providers to accept all sampling parameters.

pub fn with_allow_fallbacks(self, allow: bool) -> Self

Allow / disallow OpenRouter to use other providers as fallbacks.

pub fn with_zdr(self, zdr: bool) -> Self

Per-request Zero-Data-Retention enforcement.

pub fn with_nitro(self) -> Self

Shorthand for with_provider_sort("throughput") — equivalent to a :nitro model suffix.

pub fn with_floor(self) -> Self

Shorthand for with_provider_sort("price") — equivalent to a :floor model suffix.

pub fn with_reasoning(self, reasoning: ReasoningConfig) -> Self

Replace the full reasoning config.

pub fn with_reasoning_effort(self, effort: impl Into<String>) -> Self

Set the reasoning effort ("low", "medium", "high").

pub fn with_reasoning_max_tokens(self, max_tokens: u32) -> Self

Cap the reasoning-token budget for this request.

pub fn with_reasoning_exclude(self, exclude: bool) -> Self

Ask the provider to omit reasoning content from the response.

pub fn with_plugins(self, plugins: Vec<Plugin>) -> Self

Replace the plugin list.

pub fn with_web_search(self) -> Self

Enable the default web-search plugin. Equivalent to with_plugins(vec![Plugin::web()]), but preserves any plugins already configured (web is pushed onto the existing list).

Trait Implementations

impl Clone for ChatCompletionRequest

fn clone(&self) -> ChatCompletionRequest

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for ChatCompletionRequest

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for ChatCompletionRequest

fn default() -> ChatCompletionRequest

Returns the “default value” for a type.

impl<'de> Deserialize<'de> for ChatCompletionRequest

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for ChatCompletionRequest

fn eq(&self, other: &ChatCompletionRequest) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for ChatCompletionRequest

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for ChatCompletionRequest