How does Python 3.13's free-threaded mode achieve true multi-core parallelism without the Global Interpreter Lock (GIL)?

# Install free-threaded Python
# uv python install 3.13t

import sysconfig
import threading
import time

print(f"GIL: {sysconfig.get_config_var('Py_GIL_DISABLED')}")  # 1 = free-threaded

def cpu_work():
    total = 0
    for i in range(10**8):  # CPU intensive
        total += i * i
    return total

# 8 threads true parallelism (no GIL bottleneck)
start = time.time()
threads = [threading.Thread(target=cpu_work) for _ in range(8)]
for t in threads: t.start()
for t in threads: t.join()
print(f"Free-threaded: {time.time() - start:.2f}s")  # ~0.5s vs 4s GIL