diff --git a/docs/source/models/supported_models.rst b/docs/source/models/supported_models.rst
index 55835d945b00c..87f45cf695c8d 100644
--- a/docs/source/models/supported_models.rst
+++ b/docs/source/models/supported_models.rst
@@ -333,7 +333,7 @@ Text Embedding
   * - :code:`MistralModel`
     - Mistral-based
     - :code:`intfloat/e5-mistral-7b-instruct`, etc.
-    - 
+    - ✅︎
     - ✅︎
 
 .. important::
diff --git a/vllm/model_executor/models/llama.py b/vllm/model_executor/models/llama.py
index 6c0a8b5ef8451..d768a57b7ef8a 100644
--- a/vllm/model_executor/models/llama.py
+++ b/vllm/model_executor/models/llama.py
@@ -627,7 +627,7 @@ def permute(w: torch.Tensor, n_heads: int):
         return name, loaded_weight
 
 
-class LlamaEmbeddingModel(nn.Module, SupportsPP):
+class LlamaEmbeddingModel(nn.Module, SupportsLoRA, SupportsPP):
     """
     A model that uses Llama with additional embedding functionalities.
 
@@ -638,6 +638,19 @@ class LlamaEmbeddingModel(nn.Module, SupportsPP):
         model: An instance of LlamaModel used for forward operations.
         _pooler: An instance of Pooler used for pooling operations.
     """
+    packed_modules_mapping = {
+        "qkv_proj": ["q_proj", "k_proj", "v_proj"],
+        "gate_up_proj": ["gate_proj", "up_proj"]
+    }
+
+    # LoRA specific attributes
+    supported_lora_modules = [
+        "qkv_proj", "o_proj", "gate_up_proj", "down_proj", "embed_tokens"
+    ]
+    embedding_modules = {
+        "embed_tokens": "input_embeddings",
+    }
+    embedding_padding_modules = []
 
     def __init__(
         self,
@@ -679,3 +692,8 @@ def load_weights(self, weights: Iterable[Tuple[str, torch.Tensor]]):
 
     def load_kv_cache_scales(self, quantization_param_path: str) -> None:
         self.model.load_kv_cache_scales(quantization_param_path)
+
+    # LRUCacheWorkerLoRAManager instantiation requires model config.
+    @property
+    def config(self):
+        return self.model.config