diff --git a/.changelog/5288.changed b/.changelog/5288.changed
new file mode 100644
index 0000000000..224761ac36
--- /dev/null
+++ b/.changelog/5288.changed
@@ -0,0 +1 @@
+opentelemetry-api: remove unnecessary copy in iterator
\ No newline at end of file
diff --git a/opentelemetry-api/src/opentelemetry/attributes/__init__.py b/opentelemetry-api/src/opentelemetry/attributes/__init__.py
index 58b2901a6b..4669468b8e 100644
--- a/opentelemetry-api/src/opentelemetry/attributes/__init__.py
+++ b/opentelemetry-api/src/opentelemetry/attributes/__init__.py
@@ -309,7 +309,7 @@ def __delitem__(self, key: str) -> None:
 
     def __iter__(self):  # type: ignore
         with self._lock:
-            return iter(self._dict.copy())  # type: ignore
+            return iter(list(self._dict))  # type: ignore
 
     def __len__(self) -> int:
         return len(self._dict)
diff --git a/opentelemetry-sdk/benchmarks/trace/test_benchmark_trace.py b/opentelemetry-sdk/benchmarks/trace/test_benchmark_trace.py
index 7757dbab12..0e110514ba 100644
--- a/opentelemetry-sdk/benchmarks/trace/test_benchmark_trace.py
+++ b/opentelemetry-sdk/benchmarks/trace/test_benchmark_trace.py
@@ -1,10 +1,12 @@
 # Copyright The OpenTelemetry Authors
 # SPDX-License-Identifier: Apache-2.0
 
+import tracemalloc
 from functools import lru_cache
 
 import pytest
 
+from opentelemetry.attributes import BoundedAttributes
 from opentelemetry.sdk.resources import Resource
 from opentelemetry.sdk.trace import (
     ReadableSpan,
@@ -192,3 +194,24 @@ def benchmark_read_links():
         span.end()
 
     benchmark(benchmark_read_links)
+
+
+@pytest.mark.parametrize("num_attrs", [1, 10, 50, 128])
+def test_bounded_attribute_iterator(benchmark, num_attrs):
+    attrs = BoundedAttributes(
+        attributes={f"key{i}": f"value{i}" for i in range(num_attrs)}
+    )
+
+    peaks = []
+    for _ in range(200):
+        tracemalloc.start()
+        list(attrs)
+        _, peak = tracemalloc.get_traced_memory()
+        tracemalloc.stop()
+        peaks.append(peak)
+    benchmark.extra_info["mean_alloc_bytes"] = sum(peaks) / len(peaks)
+
+    def benchmark_iter():
+        list(attrs)
+
+    benchmark(benchmark_iter)