Python 3.12: What's New, Release History & EOL Status

What Is New in Python 3.12

Python 3.12 continues the performance work from 3.11, adding a new type parameter syntax for generics, f-string parser rewrite that unlocks previously invalid syntax, Linux perf profiler integration, and the largest stdlib cleanup in Python's history via PEP 594 removals. Sub-interpreters gain their own GIL, moving toward genuine interpreter-level isolation.

What Is the New Type Parameter Syntax in Python 3.12?

PEP 695 introduces a compact syntax for defining generics directly in def, class, and the new type statement. Before 3.12, defining a generic function required importing TypeVar and declaring it separately -- now it is inline.

# Before 3.12
from typing import TypeVar
T = TypeVar("T")
def first(items: list[T]) -> T:
    return items[0]

# Python 3.12+
def first[T](items: list[T]) -> T:
    return items[0]

# Generic class
class Stack[T]:
    def __init__(self) -> None:
        self._items: list[T] = []
    def push(self, item: T) -> None:
        self._items.append(item)
    def pop(self) -> T:
        return self._items.pop()

# Type alias
type Vector = list[float]
type Matrix[T] = list[list[T]]

The type statement creates a TypeAliasType object, which type checkers understand as a true alias rather than a variable assignment. This resolves an old ambiguity where MyList = list[str] looked like a regular assignment.

What Changed About F-Strings in Python 3.12?

The f-string parser was completely rewritten (PEP 701). The old parser was a hack bolted onto the tokenizer and had arbitrary restrictions. The new parser allows:

• Reuse of any quote type inside an expression: f"{'nested string'}"
• Nested f-strings: f"{f"{f"{x}"}"}"
• Multiline expressions inside {} with comments
• Backslashes inside expression blocks (previously a SyntaxError)

# All these now work:
names = ["Alice", "Bob"]
print(f"{", ".join(names)}")          # nested same-quote string
print(f"{"\n".join(names)}")          # backslash in expression
total = f"{
    sum(range(100))   # multiline with comment
}")
print(f"nested: {f"inner {1 + 1}"}")

How Does Linux perf Profiler Support Work in 3.12?

Running python -X perf script.py generates profiling data readable by the Linux perf tool. CPython emits DWARF frame information so that perf report shows Python function names alongside C stack frames -- giving a unified view of where time is spent across both layers.

# Profile with Linux perf
python -X perf -m my_app

# Then analyze
perf report

This is aimed at production profiling scenarios where you need to see the interaction between Python code and C extensions or system calls -- situations where cProfile alone is insufficient.

What Was Removed in Python 3.12?

Python 3.12 completed the removal of modules deprecated in 3.11 and earlier:

• distutils -- use setuptools (PyPI) or the build package instead.
• asynchat, asyncore -- use asyncio.
• smtpd -- use aiosmtpd or a dedicated SMTP library.
• lib2to3 and the 2to3 tool -- Python 2 migration tools are obsolete.
• imp module -- use importlib.
• pkgutil.find_loader() -- use importlib.util.find_spec().

If your code imports any of these, it will raise ModuleNotFoundError on Python 3.12. Migration is straightforward in all cases.

What New Standard Library Functions Shipped in 3.12?

• pathlib.Path.walk() -- like os.walk() but returns Path objects; supports top_down and on_error.
• itertools.batched(iterable, n) -- yields tuples of up to n items; clean chunking without manual slicing.
• os.process_cpu_count() -- counts CPUs available to the process (respects cgroups), unlike os.cpu_count() which counts all physical CPUs.
• sys.monitoring -- low-overhead event monitoring for debuggers and profilers to hook into without patching bytecode.
• dataclasses.KW_ONLY -- marker to make all subsequent fields keyword-only in a dataclass definition.