diff --git a/learning-examples/get_graduating_tokens.py b/learning-examples/get_graduating_tokens.py
new file mode 100644
index 0000000..1695c13
--- /dev/null
+++ b/learning-examples/get_graduating_tokens.py
@@ -0,0 +1,84 @@
+import asyncio
+import os
+import struct
+
+from dotenv import load_dotenv
+from solana.rpc.async_api import AsyncClient
+from solana.rpc.types import MemcmpOpts
+from solders.pubkey import Pubkey
+
+load_dotenv()
+
+RPC_ENDPOINT = os.environ.get("SOLANA_NODE_RPC_ENDPOINT1")
+PUMP_PROGRAM_ID = Pubkey.from_string("6EF8rrecthR5Dkzon8Nwu78hRvfCKubJ14M5uBEwF6P")
+# Known 8-byte discriminator for bonding curve accounts
+# This ensures we're only fetching bonding curve accounts
+BONDING_CURVE_DISCRIMINATOR_BYTES = bytes.fromhex("17b7f83760d8ac60")
+
+
+async def get_bonding_curves_by_reserves():
+    """
+    Fetch bonding curve accounts from the Pump.fun program that have real_token_reserves
+    below a defined threshold and are not marked as complete (i.e., not migrated).
+    """
+
+    # Define the reserve threshold we're interested in
+    # 100 trillion (in token base units)
+    threshold = 100_000_000_000_000
+    
+    # Convert the threshold into 8-byte little-endian format (as stored on-chain)
+    threshold_bytes = threshold.to_bytes(8, 'little')
+    
+    # Extract the 2 most significant bytes (bytes 6 and 7 in little-endian)
+    # This helps us pre-filter for values less than 2^48 (~281T)
+    msb_prefix = threshold_bytes[6:]
+
+    async with AsyncClient(RPC_ENDPOINT, commitment="processed", timeout=120) as client:
+        # Define on-chain filters for getProgramAccounts
+        filters = [
+            MemcmpOpts(offset=0, bytes=BONDING_CURVE_DISCRIMINATOR_BYTES),  # only bonding curve accounts
+            MemcmpOpts(offset=30, bytes=msb_prefix),          # real_token_reserves < ~281T
+            MemcmpOpts(offset=48, bytes=b'\x00'),             # complete == False (not migrated)
+        ]
+
+        # Query accounts matching filters
+        response = await client.get_program_accounts(
+            PUMP_PROGRAM_ID,
+            encoding="base64",
+            filters=filters
+        )
+
+        result = []
+        for acc in response.value:
+            raw = acc.account.data
+
+            # Parse the account layout according to known struct (49 bytes):
+            # [discriminator][virtual_token_reserves][virtual_sol_reserves]
+            # [real_token_reserves][real_sol_reserves][token_total_supply][complete]
+            # Example: https://solscan.io/account/ASwuqjAGjWVhojrdot9yrnXTV4hMdTEryAWkGn4UUJma#anchorData
+
+            offset = 8  # skip 8-byte discriminator
+            offset += 8  # skip virtual_token_reserves
+            offset += 8  # skip virtual_sol_reserves
+
+            # Extract real_token_reserves (u64 = 8 bytes, little-endian)
+            real_token_reserves = struct.unpack("<Q", raw[offset:offset + 8])[0]
+            offset += 8
+
+            # Post-filter: ensure value is really below the defined threshold
+            if real_token_reserves < threshold:
+                print(f"Pubkey: {acc.pubkey}")
+                print(f"Real token reserves: {real_token_reserves / 10**6} tokens")
+                print("="*50)
+                result.append(acc)
+
+        return result
+
+
+async def main():
+    bonding_curves_response = await get_bonding_curves_by_reserves()
+    print(f"Total matches: {len(bonding_curves_response)}")
+
+
+if __name__ == "__main__":
+    asyncio.run(main())