diff --git a/src/llm/anthropic.rs b/src/llm/anthropic.rs
index 57d968aa..08315523 100644
--- a/src/llm/anthropic.rs
+++ b/src/llm/anthropic.rs
@@ -1,6 +1,5 @@
 use crate::llm::{
-    LlmGenerateRequest, LlmGenerateResponse, LlmGenerationClient, LlmSpec, OutputFormat,
-    ToJsonSchemaOptions,
+    LlmClient, LlmGenerateRequest, LlmGenerateResponse, OutputFormat, ToJsonSchemaOptions,
 };
 use anyhow::{Context, Result, bail};
 use async_trait::async_trait;
@@ -11,19 +10,20 @@ use crate::api_bail;
 use urlencoding::encode;
 
 pub struct Client {
-    model: String,
     api_key: String,
     client: reqwest::Client,
 }
 
 impl Client {
-    pub async fn new(spec: LlmSpec) -> Result<Self> {
+    pub async fn new(address: Option<String>) -> Result<Self> {
+        if address.is_some() {
+            api_bail!("Anthropic doesn't support custom API address");
+        }
         let api_key = match std::env::var("ANTHROPIC_API_KEY") {
             Ok(val) => val,
             Err(_) => api_bail!("ANTHROPIC_API_KEY environment variable must be set"),
         };
         Ok(Self {
-            model: spec.model,
             api_key,
             client: reqwest::Client::new(),
         })
@@ -31,7 +31,7 @@ impl Client {
 }
 
 #[async_trait]
-impl LlmGenerationClient for Client {
+impl LlmClient for Client {
     async fn generate<'req>(
         &self,
         request: LlmGenerateRequest<'req>,
@@ -42,7 +42,7 @@ impl LlmGenerationClient for Client {
         })];
 
         let mut payload = serde_json::json!({
-            "model": self.model,
+            "model": request.model,
             "messages": messages,
             "max_tokens": 4096
         });
diff --git a/src/llm/gemini.rs b/src/llm/gemini.rs
index f48f839d..09c6b8ca 100644
--- a/src/llm/gemini.rs
+++ b/src/llm/gemini.rs
@@ -1,7 +1,6 @@
 use crate::api_bail;
 use crate::llm::{
-    LlmGenerateRequest, LlmGenerateResponse, LlmGenerationClient, LlmSpec, OutputFormat,
-    ToJsonSchemaOptions,
+    LlmClient, LlmGenerateRequest, LlmGenerateResponse, OutputFormat, ToJsonSchemaOptions,
 };
 use anyhow::{Context, Result, bail};
 use async_trait::async_trait;
@@ -9,19 +8,20 @@ use serde_json::Value;
 use urlencoding::encode;
 
 pub struct Client {
-    model: String,
     api_key: String,
     client: reqwest::Client,
 }
 
 impl Client {
-    pub async fn new(spec: LlmSpec) -> Result<Self> {
+    pub async fn new(address: Option<String>) -> Result<Self> {
+        if address.is_some() {
+            api_bail!("Gemini doesn't support custom API address");
+        }
         let api_key = match std::env::var("GEMINI_API_KEY") {
             Ok(val) => val,
             Err(_) => api_bail!("GEMINI_API_KEY environment variable must be set"),
         };
         Ok(Self {
-            model: spec.model,
             api_key,
             client: reqwest::Client::new(),
         })
@@ -47,7 +47,7 @@ fn remove_additional_properties(value: &mut Value) {
 }
 
 #[async_trait]
-impl LlmGenerationClient for Client {
+impl LlmClient for Client {
     async fn generate<'req>(
         &self,
         request: LlmGenerateRequest<'req>,
@@ -79,7 +79,7 @@ impl LlmGenerationClient for Client {
         let api_key = &self.api_key;
         let url = format!(
             "https://generativelanguage.googleapis.com/v1beta/models/{}:generateContent?key={}",
-            encode(&self.model),
+            encode(request.model),
             encode(api_key)
         );
 
diff --git a/src/llm/litellm.rs b/src/llm/litellm.rs
index 27648747..85d1b50e 100644
--- a/src/llm/litellm.rs
+++ b/src/llm/litellm.rs
@@ -4,19 +4,13 @@ use async_openai::config::OpenAIConfig;
 pub use super::openai::Client;
 
 impl Client {
-    pub async fn new_litellm(spec: super::LlmSpec) -> anyhow::Result<Self> {
-        let address = spec
-            .address
-            .clone()
-            .unwrap_or_else(|| "http://127.0.0.1:4000".to_string());
+    pub async fn new_litellm(address: Option<String>) -> anyhow::Result<Self> {
+        let address = address.unwrap_or_else(|| "http://127.0.0.1:4000".to_string());
         let api_key = std::env::var("LITELLM_API_KEY").ok();
         let mut config = OpenAIConfig::new().with_api_base(address);
         if let Some(api_key) = api_key {
             config = config.with_api_key(api_key);
         }
-        Ok(Client::from_parts(
-            OpenAIClient::with_config(config),
-            spec.model,
-        ))
+        Ok(Client::from_parts(OpenAIClient::with_config(config)))
     }
 }
diff --git a/src/llm/mod.rs b/src/llm/mod.rs
index ea4aa58e..32bdd5a1 100644
--- a/src/llm/mod.rs
+++ b/src/llm/mod.rs
@@ -19,9 +19,9 @@ pub enum LlmApiType {
 
 #[derive(Debug, Clone, Serialize, Deserialize)]
 pub struct LlmSpec {
-    api_type: LlmApiType,
-    address: Option<String>,
-    model: String,
+    pub api_type: LlmApiType,
+    pub address: Option<String>,
+    pub model: String,
 }
 
 #[derive(Debug)]
@@ -34,6 +34,7 @@ pub enum OutputFormat<'a> {
 
 #[derive(Debug)]
 pub struct LlmGenerateRequest<'a> {
+    pub model: &'a str,
     pub system_prompt: Option<Cow<'a, str>>,
     pub user_prompt: Cow<'a, str>,
     pub output_format: Option<OutputFormat<'a>>,
@@ -45,7 +46,7 @@ pub struct LlmGenerateResponse {
 }
 
 #[async_trait]
-pub trait LlmGenerationClient: Send + Sync {
+pub trait LlmClient: Send + Sync {
     async fn generate<'req>(
         &self,
         request: LlmGenerateRequest<'req>,
@@ -61,25 +62,23 @@ mod ollama;
 mod openai;
 mod openrouter;
 
-pub async fn new_llm_generation_client(spec: LlmSpec) -> Result<Box<dyn LlmGenerationClient>> {
-    let client = match spec.api_type {
-        LlmApiType::Ollama => {
-            Box::new(ollama::Client::new(spec).await?) as Box<dyn LlmGenerationClient>
-        }
-        LlmApiType::OpenAi => {
-            Box::new(openai::Client::new(spec).await?) as Box<dyn LlmGenerationClient>
-        }
-        LlmApiType::Gemini => {
-            Box::new(gemini::Client::new(spec).await?) as Box<dyn LlmGenerationClient>
-        }
+pub async fn new_llm_generation_client(
+    api_type: LlmApiType,
+    address: Option<String>,
+) -> Result<Box<dyn LlmClient>> {
+    let client = match api_type {
+        LlmApiType::Ollama => Box::new(ollama::Client::new(address).await?) as Box<dyn LlmClient>,
+        LlmApiType::OpenAi => Box::new(openai::Client::new(address).await?) as Box<dyn LlmClient>,
+        LlmApiType::Gemini => Box::new(gemini::Client::new(address).await?) as Box<dyn LlmClient>,
         LlmApiType::Anthropic => {
-            Box::new(anthropic::Client::new(spec).await?) as Box<dyn LlmGenerationClient>
+            Box::new(anthropic::Client::new(address).await?) as Box<dyn LlmClient>
         }
         LlmApiType::LiteLlm => {
-            Box::new(litellm::Client::new_litellm(spec).await?) as Box<dyn LlmGenerationClient>
+            Box::new(litellm::Client::new_litellm(address).await?) as Box<dyn LlmClient>
+        }
+        LlmApiType::OpenRouter => {
+            Box::new(openrouter::Client::new_openrouter(address).await?) as Box<dyn LlmClient>
         }
-        LlmApiType::OpenRouter => Box::new(openrouter::Client::new_openrouter(spec).await?)
-            as Box<dyn LlmGenerationClient>,
     };
     Ok(client)
 }
diff --git a/src/llm/ollama.rs b/src/llm/ollama.rs
index f2926077..7b79293f 100644
--- a/src/llm/ollama.rs
+++ b/src/llm/ollama.rs
@@ -1,4 +1,4 @@
-use super::LlmGenerationClient;
+use super::LlmClient;
 use anyhow::Result;
 use async_trait::async_trait;
 use schemars::schema::SchemaObject;
@@ -6,7 +6,6 @@ use serde::{Deserialize, Serialize};
 
 pub struct Client {
     generate_url: String,
-    model: String,
     reqwest_client: reqwest::Client,
 }
 
@@ -33,27 +32,26 @@ struct OllamaResponse {
 const OLLAMA_DEFAULT_ADDRESS: &str = "http://localhost:11434";
 
 impl Client {
-    pub async fn new(spec: super::LlmSpec) -> Result<Self> {
-        let address = match &spec.address {
+    pub async fn new(address: Option<String>) -> Result<Self> {
+        let address = match &address {
             Some(addr) => addr.trim_end_matches('/'),
             None => OLLAMA_DEFAULT_ADDRESS,
         };
         Ok(Self {
             generate_url: format!("{}/api/generate", address),
-            model: spec.model,
             reqwest_client: reqwest::Client::new(),
         })
     }
 }
 
 #[async_trait]
-impl LlmGenerationClient for Client {
+impl LlmClient for Client {
     async fn generate<'req>(
         &self,
         request: super::LlmGenerateRequest<'req>,
     ) -> Result<super::LlmGenerateResponse> {
         let req = OllamaRequest {
-            model: &self.model,
+            model: request.model,
             prompt: request.user_prompt.as_ref(),
             format: request.output_format.as_ref().map(
                 |super::OutputFormat::JsonSchema { schema, .. }| {
diff --git a/src/llm/openai.rs b/src/llm/openai.rs
index edbfa59f..3011ac5c 100644
--- a/src/llm/openai.rs
+++ b/src/llm/openai.rs
@@ -1,6 +1,6 @@
 use crate::api_bail;
 
-use super::LlmGenerationClient;
+use super::LlmClient;
 use anyhow::Result;
 use async_openai::{
     Client as OpenAIClient,
@@ -16,16 +16,15 @@ use async_trait::async_trait;
 
 pub struct Client {
     client: async_openai::Client<OpenAIConfig>,
-    model: String,
 }
 
 impl Client {
-    pub(crate) fn from_parts(client: async_openai::Client<OpenAIConfig>, model: String) -> Self {
-        Self { client, model }
+    pub(crate) fn from_parts(client: async_openai::Client<OpenAIConfig>) -> Self {
+        Self { client }
     }
 
-    pub async fn new(spec: super::LlmSpec) -> Result<Self> {
-        if let Some(address) = spec.address {
+    pub async fn new(address: Option<String>) -> Result<Self> {
+        if let Some(address) = address {
             api_bail!("OpenAI doesn't support custom API address: {address}");
         }
         // Verify API key is set
@@ -35,13 +34,12 @@ impl Client {
         Ok(Self {
             // OpenAI client will use OPENAI_API_KEY env variable by default
             client: OpenAIClient::new(),
-            model: spec.model,
         })
     }
 }
 
 #[async_trait]
-impl LlmGenerationClient for Client {
+impl LlmClient for Client {
     async fn generate<'req>(
         &self,
         request: super::LlmGenerateRequest<'req>,
@@ -70,7 +68,7 @@ impl LlmGenerationClient for Client {
 
         // Create the chat completion request
         let request = CreateChatCompletionRequest {
-            model: self.model.clone(),
+            model: request.model.to_string(),
             messages,
             response_format: match request.output_format {
                 Some(super::OutputFormat::JsonSchema { name, schema }) => {
diff --git a/src/llm/openrouter.rs b/src/llm/openrouter.rs
index cb775788..ecf4d0fa 100644
--- a/src/llm/openrouter.rs
+++ b/src/llm/openrouter.rs
@@ -4,19 +4,13 @@ use async_openai::config::OpenAIConfig;
 pub use super::openai::Client;
 
 impl Client {
-    pub async fn new_openrouter(spec: super::LlmSpec) -> anyhow::Result<Self> {
-        let address = spec
-            .address
-            .clone()
-            .unwrap_or_else(|| "https://openrouter.ai/api/v1".to_string());
+    pub async fn new_openrouter(address: Option<String>) -> anyhow::Result<Self> {
+        let address = address.unwrap_or_else(|| "https://openrouter.ai/api/v1".to_string());
         let api_key = std::env::var("OPENROUTER_API_KEY").ok();
         let mut config = OpenAIConfig::new().with_api_base(address);
         if let Some(api_key) = api_key {
             config = config.with_api_key(api_key);
         }
-        Ok(Client::from_parts(
-            OpenAIClient::with_config(config),
-            spec.model,
-        ))
+        Ok(Client::from_parts(OpenAIClient::with_config(config)))
     }
 }
diff --git a/src/ops/functions/extract_by_llm.rs b/src/ops/functions/extract_by_llm.rs
index 904bb234..ec9782c8 100644
--- a/src/ops/functions/extract_by_llm.rs
+++ b/src/ops/functions/extract_by_llm.rs
@@ -1,8 +1,6 @@
 use crate::prelude::*;
 
-use crate::llm::{
-    LlmGenerateRequest, LlmGenerationClient, LlmSpec, OutputFormat, new_llm_generation_client,
-};
+use crate::llm::{LlmClient, LlmGenerateRequest, LlmSpec, OutputFormat, new_llm_generation_client};
 use crate::ops::sdk::*;
 use base::json_schema::build_json_schema;
 use schemars::schema::SchemaObject;
@@ -21,7 +19,8 @@ pub struct Args {
 
 struct Executor {
     args: Args,
-    client: Box<dyn LlmGenerationClient>,
+    client: Box<dyn LlmClient>,
+    model: String,
     output_json_schema: SchemaObject,
     system_prompt: String,
     value_extractor: base::json_schema::ValueExtractor,
@@ -50,11 +49,13 @@ Output only the JSON without any additional messages or explanations."
 
 impl Executor {
     async fn new(spec: Spec, args: Args) -> Result<Self> {
-        let client = new_llm_generation_client(spec.llm_spec).await?;
+        let client =
+            new_llm_generation_client(spec.llm_spec.api_type, spec.llm_spec.address).await?;
         let schema_output = build_json_schema(spec.output_type, client.json_schema_options())?;
         Ok(Self {
             args,
             client,
+            model: spec.llm_spec.model,
             output_json_schema: schema_output.schema,
             system_prompt: get_system_prompt(&spec.instruction, schema_output.extra_instructions),
             value_extractor: schema_output.value_extractor,
@@ -75,6 +76,7 @@ impl SimpleFunctionExecutor for Executor {
     async fn evaluate(&self, input: Vec<Value>) -> Result<Value> {
         let text = self.args.text.value(&input)?.as_str()?;
         let req = LlmGenerateRequest {
+            model: &self.model,
             system_prompt: Some(Cow::Borrowed(&self.system_prompt)),
             user_prompt: Cow::Borrowed(text),
             output_format: Some(OutputFormat::JsonSchema {