Pydantic Performance issues related to schema building (Fix)

The Fix

pip install pydantic==2.11.0

Based on closed pydantic/pydantic issue #10297 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -1,19 +1,20 @@
 from __future__ import annotations
 
+import inspect
 import os
-from collections import defaultdict

repro.py

# Setup:

from typing import Union

from pydantic import create_model

simple_models = [
    create_model(
        f"SimpleModel{i}",
        **{f"f{j}": (int, ...) for j in range(10)},
    )
    for i in range(20)
]

UnionModels = Union[*simple_models]

# Statement:

Model = create_model(
    "Model",
    **{f"f{i}": (UnionModels, ...) for i in range(30)}
)

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==2.11.0\nWhen NOT to use: Avoid this fix if the schema generation logic is expected to change frequently.\n\n

Why This Fix Works in Production

Trigger: Performance issues related to schema building
Mechanism: Performance issues arise from repeated schema building without caching mechanisms
Why the fix works: Refactor and optimize schema cleaning logic to improve performance during schema generation. (first fixed release: 2.11.0).

Why This Breaks in Prod

Performance issues arise from repeated schema building without caching mechanisms
Production symptom (often without a traceback): Performance issues related to schema building

Proof / Evidence

GitHub issue: #10297
Fix PR: https://github.com/pydantic/pydantic/pull/11244
First fixed release: 2.11.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.62

Discussion

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

Jump to Sources Open on GitHub

“This is a completely private API that is not intended to be used outside Pydantic. As I mentioned, we are in the process of designing…”

@Viicos · 2025-01-14 · source

“Also strongly related: https://github.com/pydantic/pydantic/issues/10036 and https://github.com/pydantic/pydantic/issues/10074 Given this, and my familiarity with the above code, I'd love to take a stab at a metadata refactor,…”

@sydney-runkle · 2024-09-04 · source

“Though the performance benefits might not be significant, I'd argue https://github.com/pydantic/pydantic/issues/10299 is related :). The less fluff we can have in that process, the better.”

@sydney-runkle · 2024-09-04 · source

“Would it be possible to pre-build everything, say as a JSON (with an option to compress), so users will deploy their app necessary artifacts for…”

@dmarxn · 2025-01-01 · source

Failure Signature (Search String)

Performance issues related to schema building
This issue gathers info about performance issues related to schema building. This strongly relates to https://github.com/pydantic/pydantic/discussions/6748. Only time performance

Copy-friendly signature

signature.txt

Failure Signature
-----------------
Performance issues related to schema building
This issue gathers info about performance issues related to schema building. This strongly relates to https://github.com/pydantic/pydantic/discussions/6748. Only time performance is in scope.

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
Performance issues related to schema building
This issue gathers info about performance issues related to schema building. This strongly relates to https://github.com/pydantic/pydantic/discussions/6748. Only time performance is in scope.

Minimal Reproduction

repro.py

# Setup:

from typing import Union

from pydantic import create_model

simple_models = [
    create_model(
        f"SimpleModel{i}",
        **{f"f{j}": (int, ...) for j in range(10)},
    )
    for i in range(20)
]

UnionModels = Union[*simple_models]

# Statement:

Model = create_model(
    "Model",
    **{f"f{i}": (UnionModels, ...) for i in range(30)}
)

What Broke

Schema generation leads to significant delays, impacting application responsiveness.

Why It Broke

Performance issues arise from repeated schema building without caching mechanisms

Fix Options (Details)

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

pip install pydantic==2.11.0

When NOT to use: Avoid this fix if the schema generation logic is expected to change frequently.

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/11244

First fixed release: 2.11.0

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

When NOT to Use This Fix

Avoid this fix if the schema generation logic is expected to change frequently.

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.
Track RSS + object counts after deployments; alert on monotonic growth and GC pressure.
Add a long-running test that repeats the failing call path and asserts stable memory.

Version Compatibility Table

Version	Status
2.11.0	Fixed

Related Issues

No related fixes found.

Cluster: pydantic:openapi-schema Pydantic hub Pydantic best practices All hubs All clusters

Related clusters: Configuration error Data consistency Serialization

Sources

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