diff --git a/crates/llm/src/lib.rs b/crates/llm/src/lib.rs
index 5ab36fa..4897e7a 100644
--- a/crates/llm/src/lib.rs
+++ b/crates/llm/src/lib.rs
@@ -9,6 +9,7 @@ use llama_cpp_2::llama_batch::LlamaBatch;
 use llama_cpp_2::model::params::LlamaModelParams;
 use llama_cpp_2::model::{AddBos, LlamaChatMessage, LlamaChatTemplate, LlamaModel};
 use llama_cpp_2::sampling::LlamaSampler;
+use magnotia_core::tuning::{inference_thread_count, Workload};
 use serde::{Deserialize, Serialize};
 
 pub mod grammars;
@@ -163,7 +164,10 @@ impl LlmEngine {
         }
 
         let n_ctx = preflight_context_window(prompt_tokens.len(), config.max_tokens)?;
-        let thread_count = i32::try_from(magnotia_core::constants::inference_thread_count())
+        let use_gpu = self.loaded_model().map(|s| s.use_gpu).unwrap_or(false);
+        let gpu_layers = if use_gpu { u32::MAX } else { 0u32 };
+        let gpu_offloaded = use_gpu && gpu_layers >= model.n_layer();
+        let thread_count = i32::try_from(inference_thread_count(Workload::Llm, gpu_offloaded))
             .unwrap_or(4);
         let ctx_params = LlamaContextParams::default()
             .with_n_ctx(Some(