diff --git a/backend/llm.py b/backend/llm.py
index dba5720..a0f49e4 100644
--- a/backend/llm.py
+++ b/backend/llm.py
@@ -8,6 +8,7 @@ from utils import pprint_prompt
 
 
 # Actual model versions that are passed to the LLMs and stored in our logs
+# Keep in sync with s2c-saas repo & DB column `llm_version`
 class Llm(Enum):
     GPT_4_VISION = "gpt-4-vision-preview"
     CLAUDE_3_SONNET = "claude-3-sonnet-20240229"
diff --git a/backend/routes/generate_code.py b/backend/routes/generate_code.py
index 092b8d1..feb4664 100644
--- a/backend/routes/generate_code.py
+++ b/backend/routes/generate_code.py
@@ -343,10 +343,12 @@ async def stream_code(websocket: WebSocket):
     if IS_PROD:
         # Catch any errors from sending to SaaS backend and continue
         try:
+            assert exact_llm_version is not None, "exact_llm_version is not set"
             await send_to_saas_backend(
                 prompt_messages,
                 completion,
                 payment_method=payment_method,
+                llm_version=exact_llm_version,
                 auth_token=params["authToken"],
             )
         except Exception as e:
diff --git a/backend/routes/logging_utils.py b/backend/routes/logging_utils.py
index 80e0f25..bf2cdad 100644
--- a/backend/routes/logging_utils.py
+++ b/backend/routes/logging_utils.py
@@ -5,6 +5,7 @@ from typing import List
 import json
 
 from config import BACKEND_SAAS_URL, IS_PROD
+from llm import Llm
 
 
 class PaymentMethod(Enum):
@@ -19,6 +20,7 @@ async def send_to_saas_backend(
     prompt_messages: List[ChatCompletionMessageParam],
     completion: str,
     payment_method: PaymentMethod,
+    llm_version: Llm,
     auth_token: str | None = None,
 ):
     if IS_PROD:
@@ -30,6 +32,7 @@ async def send_to_saas_backend(
                     "prompt": json.dumps(prompt_messages),
                     "completion": completion,
                     "payment_method": payment_method.value,
+                    "llm_version": llm_version.value,
                 }
             )