Pydantic Change in GetCoreSchemaHandler behaviour in 2.11

The Fix

pip install pydantic==2.10.6

Based on closed pydantic/pydantic issue #11923 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -967,51 +967,6 @@ except ValidationError as e:
 ```
 
-As seen above, annotating a field with a `BaseModel` type can be used to modify or override the generated json schema.
-However, if you want to take advantage of storing metadata via `Annotated`, but you don't want to override the generated JSON
-schema, you can use the following approach with a no-op version of `__get_pydantic_core_schema__` implemented on the

repro.py

from typing import TypeVar, Generic, get_args, GenericAlias
from pydantic import TypeAdapter, GetCoreSchemaHandler
from pydantic.dataclasses import dataclass
from pydantic_core.core_schema import tagged_union_schema
from pydantic._internal._core_utils import pretty_print_core_schema

T = TypeVar('T')
U = TypeVar('U')

class Parent(Generic[T]):
    subs: list[type] = []

    @classmethod
    def __init_subclass__(cls):
        super(Parent, cls).__init_subclass__()
        Parent.subs.append(cls)

    @classmethod
    def __get_pydantic_core_schema__(cls: type, actual: type, handler: GetCoreSchemaHandler):
        if cls is Parent:
            args = get_args(actual)
            sub_types = tuple(GenericAlias(sub, args) for sub in cls.subs)
            pretty_print_core_schema(handler(sub_types[0]))
            return tagged_union_schema(
                    {t.__name__: handler(t) for t in sub_types},
                    discriminator="type",
                    ref="Parent"
                    )
        else:
            return handler(cls)


@dataclass
class C1(Parent[U]):
    c1: U

@dataclass
class C2(Parent[U]):
    c2: U

a = TypeAdapter(Parent[int]).validate_python({"type": "C1", "c1": "42"})

# Assertion passes in 2.10 but fails in 2.11 as c1 is the string "42" instead of an int
assert a.c1 == 42

verify

Re-run: uv run --with 'pydantic<2.11' --with rich demo.py
Re-run: uv run --with 'pydantic>=2.11' --with rich demo.py

fix.md

Option A — Upgrade to fixed release\npip install pydantic==2.10.6\nWhen NOT to use: This fix should not be applied if the existing behavior is required for backward compatibility.\n\n

Why This Fix Works in Production

Trigger: AssertionError
Mechanism: Optimizes calls to `get_type_ref` to reduce unnecessary function calls during schema generation, addressing issues with schema generation in Pydantic V2.11.
Why the fix works: Optimizes calls to `get_type_ref` to reduce unnecessary function calls during schema generation, addressing issues with schema generation in Pydantic V2.11. (first fixed release: 2.10.6).

Production impact:

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

Why This Breaks in Prod

Triggered by an upgrade/regression window: 2.11 breaks; 2.10.6 is the first fixed release.
Shows up under Python 3.12 in real deployments (not just unit tests).
Surfaces as: $ uv run --with 'pydantic<2.11' --with rich demo.py

Proof / Evidence

GitHub issue: #11923
Fix PR: https://github.com/pydantic/pydantic/pull/10863
First fixed release: 2.10.6
Affected versions: 2.11
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.95
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.38

Discussion

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

Jump to Sources Open on GitHub

“Thank you for the quick response and fix.”

@tpoliaw · 2025-05-28 · source

“I'm not entirely sure why https://github.com/pydantic/pydantic/pull/10863 changed this, but anyway you can fix this by properly parameterizing your subclasses”

@Viicos · 2025-05-28 · source

Failure Signature (Search String)

AssertionError

Error Message

Stack trace

error.txt

Error Message
-------------
$ uv run --with 'pydantic<2.11' --with rich demo.py
{'type': 'definition-ref', 'schema_ref': '__main__.C1:976160128[int:131761114585728]'}

$ uv run --with 'pydantic>=2.11' --with rich demo.py
{'type': 'definition-ref', 'schema_ref': '__main__.C1:261265008'}
Traceback (most recent call last):
  File "/dls/athena/scanspec/basic.py", line 43, in <module>
    assert a.c1 == 42
           ^^^^^^^^^^
AssertionError

Minimal Reproduction

repro.py

from typing import TypeVar, Generic, get_args, GenericAlias
from pydantic import TypeAdapter, GetCoreSchemaHandler
from pydantic.dataclasses import dataclass
from pydantic_core.core_schema import tagged_union_schema
from pydantic._internal._core_utils import pretty_print_core_schema

T = TypeVar('T')
U = TypeVar('U')

class Parent(Generic[T]):
    subs: list[type] = []

    @classmethod
    def __init_subclass__(cls):
        super(Parent, cls).__init_subclass__()
        Parent.subs.append(cls)

    @classmethod
    def __get_pydantic_core_schema__(cls: type, actual: type, handler: GetCoreSchemaHandler):
        if cls is Parent:
            args = get_args(actual)
            sub_types = tuple(GenericAlias(sub, args) for sub in cls.subs)
            pretty_print_core_schema(handler(sub_types[0]))
            return tagged_union_schema(
                    {t.__name__: handler(t) for t in sub_types},
                    discriminator="type",
                    ref="Parent"
                    )
        else:
            return handler(cls)


@dataclass
class C1(Parent[U]):
    c1: U

@dataclass
class C2(Parent[U]):
    c2: U

a = TypeAdapter(Parent[int]).validate_python({"type": "C1", "c1": "42"})

# Assertion passes in 2.10 but fails in 2.11 as c1 is the string "42" instead of an int
assert a.c1 == 42

Environment

Python: 3.12
Pydantic: 2

What Broke

Subtypes are not deserialized correctly, leading to assertion failures in production.

Fix Options (Details)

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

pip install pydantic==2.10.6

When NOT to use: This fix should not be applied if the existing behavior is required for backward compatibility.

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

Option D — Guard side-effects with OnceOnly Guardrail for side-effects

Mitigate duplicate external side-effects under retries/timeouts/agent loops by gating the operation before calling external systems.

Place OnceOnly between your code/agent and real side-effects (Stripe, emails, CRM, APIs).
Use a stable key per side-effect (e.g., customer_id + action + idempotency_key).
Fail-safe: configure fail-open vs fail-closed based on blast radius and spend risk.
This does NOT fix data corruption; it only prevents duplicate side-effects.

Show example snippet (optional)

onceonly.py

from onceonly import OnceOnly
import os

once = OnceOnly(api_key=os.environ["ONCEONLY_API_KEY"], fail_open=True)

# Stable idempotency key per real side-effect.
# Use a request id / job id / webhook delivery id / Stripe event id, etc.
event_id = "evt_..."  # replace
key = f"stripe:webhook:{event_id}"

res = once.check_lock(key=key, ttl=3600)
if res.duplicate:
    return {"status": "already_processed"}

# Safe to execute the side-effect exactly once.
handle_event(event_id)

See OnceOnly SDK

When NOT to use: Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

Fix reference: https://github.com/pydantic/pydantic/pull/10863

First fixed release: 2.10.6

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

When NOT to Use This Fix

This fix should not be applied if the existing behavior is required for backward compatibility.
Do not use this to hide logic bugs or data corruption. Use it to block duplicate external side-effects and enforce tool permissions/spend caps.

Verify Fix

verify

Re-run: uv run --with 'pydantic<2.11' --with rich demo.py
Re-run: uv run --with 'pydantic>=2.11' --with rich demo.py

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
2.11	Broken
2.10.6	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 Change in GetCoreSchemaHandler behaviour in 2.11 (Fix)