Pydantic pydantic using abc which can have memory and

The Fix

pip install pydantic==2.12.0

Based on closed pydantic/pydantic issue #11605 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -10,7 +10,7 @@
 from functools import cache, partial, wraps
 from types import FunctionType
-from typing import TYPE_CHECKING, Any, Callable, Generic, Literal, NoReturn, cast
+from typing import TYPE_CHECKING, Any, Callable, Generic, Literal, NoReturn, TypeVar, cast

repro.py

# https://github.com/peterchenadded/pydantic-abc-performance

from pydantic import BaseModel
import logging
import time

class MyEntity1Pattern(BaseModel):
    pass

class MyEntity2Pattern(BaseModel):
    pass

  
def _get_my_entity1_pattern():
    class MyTest1(MyEntity1Pattern):
        pass
    
    return MyTest1
  
def _get_my_entity2_pattern():
    class MyTest2(MyEntity2Pattern):
        pass
    
    return MyTest2

class Measurement(BaseModel):
    my_entity_1_pattern_check: float = 0
    my_entity_2_pattern_check: float = 0
    success: int = 0
    failed: int = 0

    def count(self, input: bool) -> None:
        if input:
            self.success += 1
        else:
            self.failed += 1
    
def test_performance():
    my_objects = []
    logging.info("Started getting objects to check")
    for i in range(7000):
        if i % 2 == 0:
            my_objects.append(_get_my_entity1_pattern()())
        else:
            my_objects.append(_get_my_entity2_pattern()())
    
    logging.info("Completed len(my_objects) = %d", len(my_objects)) 
    logging.info("Started checking objects")

    measurement = Measurement()
    for m in my_objects:
        start = time.time()
        measurement.count(isinstance(m, MyEntity1Pattern))
        measurement.my_entity_1_pattern_check += time.time() - start

        start = time.time()
        measurement.count(isinstance(m, MyEntity2Pattern))
        measurement.my_entity_2_pattern_check += time.time() - start

    logging.info("Completed checking classes")
    logging.info("measures:\n%s", measurement.model_dump_json(indent=4))

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

fix.md

Option A — Upgrade to fixed release\npip install pydantic==2.12.0\nWhen NOT to use: This fix is not suitable if virtual subclasses are essential for your application.\n\n

Why This Fix Works in Production

Trigger: It took 1 second to isinstance 7000 distinct class objects on my custom changes in test_performance_fix.py with --profile.
Mechanism: Pydantic's use of abc leads to excessive memory usage during isinstance checks with many distinct classes
Why the fix works: Warns if registering virtual subclasses on Pydantic models to avoid memory and performance issues. (first fixed release: 2.12.0).

Production impact:

If left unfixed, this can cause silent data inconsistencies that propagate (bad cache entries, incorrect downstream decisions).

Why This Breaks in Prod

Pydantic's use of abc leads to excessive memory usage during isinstance checks with many distinct classes
Production symptom (often without a traceback): It took 1 second to isinstance 7000 distinct class objects on my custom changes in test_performance_fix.py with --profile.

Proof / Evidence

GitHub issue: #11605
Fix PR: https://github.com/pydantic/pydantic/pull/11669
First fixed release: 2.12.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.39

Discussion

High-signal excerpts from the issue thread (symptoms, repros, edge-cases).

Jump to Sources Open on GitHub

“> If someone relied on .register() behavior, what should used to use it on 2.12? If someone does, it isn't working anyway, for at least…”

@Viicos · 2025-04-29 · confirmation · source

“> That said the 2.11 patch #11116 to fix the issue should then be reverted if we follow the logic”

@Viicos · 2025-04-29 · source

“PydanticAI Github Bot Found 1 issues similar to this one: 1. "https://github.com/pydantic/pydantic/issues/9458" (95% similar)”

@pydanticai-bot · 2025-03-22 · source

“@Viicos thanks for looking into it further and finding another bug”

@peterchenadded · 2025-03-28 · source

Failure Signature (Search String)

It took 1 second to isinstance 7000 distinct class objects on my custom changes in test_performance_fix.py with --profile.
If I changed it to 10_000 distinct class objects it gets killed on my machine.

Copy-friendly signature

signature.txt

Failure Signature
-----------------
It took 1 second to isinstance 7000 distinct class objects on my custom changes in test_performance_fix.py with --profile.
If I changed it to 10_000 distinct class objects it gets killed on my machine.

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
It took 1 second to isinstance 7000 distinct class objects on my custom changes in test_performance_fix.py with --profile.
If I changed it to 10_000 distinct class objects it gets killed on my machine.

Minimal Reproduction

repro.py

# https://github.com/peterchenadded/pydantic-abc-performance

from pydantic import BaseModel
import logging
import time

class MyEntity1Pattern(BaseModel):
    pass

class MyEntity2Pattern(BaseModel):
    pass

  
def _get_my_entity1_pattern():
    class MyTest1(MyEntity1Pattern):
        pass
    
    return MyTest1
  
def _get_my_entity2_pattern():
    class MyTest2(MyEntity2Pattern):
        pass
    
    return MyTest2

class Measurement(BaseModel):
    my_entity_1_pattern_check: float = 0
    my_entity_2_pattern_check: float = 0
    success: int = 0
    failed: int = 0

    def count(self, input: bool) -> None:
        if input:
            self.success += 1
        else:
            self.failed += 1
    
def test_performance():
    my_objects = []
    logging.info("Started getting objects to check")
    for i in range(7000):
        if i % 2 == 0:
            my_objects.append(_get_my_entity1_pattern()())
        else:
            my_objects.append(_get_my_entity2_pattern()())
    
    logging.info("Completed len(my_objects) = %d", len(my_objects)) 
    logging.info("Started checking objects")

    measurement = Measurement()
    for m in my_objects:
        start = time.time()
        measurement.count(isinstance(m, MyEntity1Pattern))
        measurement.my_entity_1_pattern_check += time.time() - start

        start = time.time()
        measurement.count(isinstance(m, MyEntity2Pattern))
        measurement.my_entity_2_pattern_check += time.time() - start

    logging.info("Completed checking classes")
    logging.info("measures:\n%s", measurement.model_dump_json(indent=4))

Environment

Pydantic: 2

What Broke

Severe memory and performance issues occur when checking types of many distinct Pydantic models.

Why It Broke

Pydantic's use of abc leads to excessive memory usage during isinstance checks with many distinct classes

Fix Options (Details)

Option A — Upgrade to fixed release Safe default (recommended)

pip install pydantic==2.12.0

When NOT to use: This fix is not suitable if virtual subclasses are essential for your application.

Use when you can deploy the upstream fix. It is usually lower-risk than long-lived workarounds.

Fix reference: https://github.com/pydantic/pydantic/pull/11669

First fixed release: 2.12.0

Last verified: 2026-02-09. Validate in your environment.

When NOT to Use This Fix

This fix is not suitable if virtual subclasses are essential for your application.

Verify Fix

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

Did This Fix Work in Your Case?

Quick signal helps us prioritize which fixes to verify and improve.

Prevention

Add a CI check that diffs key outputs after upgrades (OpenAPI schema snapshots, JSON payload shapes, CLI output).
Upgrade behind a canary and run integration tests against the canary before 100% rollout.

Version Compatibility Table

Version	Status
2.12.0	Fixed

Related Issues

No related fixes found.

Cluster: pydantic:data-consistency Pydantic hub Pydantic best practices All hubs All clusters

Related clusters: Configuration error OpenAPI schema Serialization

Sources

We don’t republish the full GitHub discussion text. Use the links above for context.

Pydantic pydantic using abc which can have memory and performance issues (Fix)