Pydantic generated json schema is non-deterministic when

The Fix

Ensures deterministic JSON schema defaults by sorting sets, addressing the issue of non-deterministic ordering in JSON schema generation when using unordered collections as defaults.

Based on closed pydantic/pydantic issue #12754 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -965,6 +965,10 @@ print(Model(set_of_ints=['1', '2', '3']).set_of_ints)
 ```
 
+<h4>JSON Schema</h4>
+
+Pydantic does best effort to sort default values that are [`collections.abc.Set`][] instances.

repro.py

import enum
import json

import pydantic


class MyEnum(str, enum.Enum):
    A = "a"
    B = "b"
    # strictly speaking, only 2 enum members are required to reproduce, but
    # it is more consistent the more members you have
    C = "c"
    D = "d"
    E = "e"
    F = "f"
    G = "g"


class MyModel(pydantic.BaseModel):
    field: frozenset[MyEnum] = frozenset(MyEnum)


schema = MyModel.model_json_schema()
print(json.dumps(schema, indent=2))

verify

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

fix.md

Option A — Apply the official fix\nEnsures deterministic JSON schema defaults by sorting sets, addressing the issue of non-deterministic ordering in JSON schema generation when using unordered collections as defaults.\nWhen NOT to use: This fix is not suitable for cases where the order of set elements is significant.\n\n

Why This Fix Works in Production

Trigger: import enum
Mechanism: The JSON schema generation fails to maintain consistent ordering for set-like defaults due to Python's hash randomization

Production impact:

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

Why This Breaks in Prod

The JSON schema generation fails to maintain consistent ordering for set-like defaults due to Python's hash randomization
Surfaces as: import enum

Proof / Evidence

GitHub issue: #12754
Fix PR: https://github.com/pydantic/pydantic/pull/12760
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.70
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.56

Discussion

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

Jump to Sources Open on GitHub

“This looks like a determinism issue in json schema generation rather than a schema design question”

@AghastyGD · 2026-01-31 · source

“I think we could go this way, but this will possibly impact existing setups, so we'll have to wait for https://github.com/pydantic/pydantic/labels/v3 if we want to…”

@Viicos · 2026-01-31 · source

“> I think we could go this way, but this will possibly impact existing setups, so we'll have to wait for v3 Under consideration for…”

@AghastyGD · 2026-01-31 · source

“> possibly impact existing setups Can you expand on this? What kind of impact are you concerned about? To me, this feels more like a…”

@cj81499 · 2026-01-31 · source

Failure Signature (Search String)

import enum

Error Message

Stack trace

error.txt

Error Message
-------------
import enum
import json

import pydantic


class MyEnum(str, enum.Enum):
    A = "a"
    B = "b"
    # strictly speaking, only 2 enum members are required to reproduce, but
    # it is more consistent the more members you have
    C = "c"
    D = "d"
    E = "e"
    F = "f"
    G = "g"


class MyModel(pydantic.BaseModel):
    field: frozenset[MyEnum] = frozenset(MyEnum)


schema = MyModel.model_json_schema()
print(json.dumps(schema, indent=2))

Minimal Reproduction

repro.py

import enum
import json

import pydantic


class MyEnum(str, enum.Enum):
    A = "a"
    B = "b"
    # strictly speaking, only 2 enum members are required to reproduce, but
    # it is more consistent the more members you have
    C = "c"
    D = "d"
    E = "e"
    F = "f"
    G = "g"


class MyModel(pydantic.BaseModel):
    field: frozenset[MyEnum] = frozenset(MyEnum)


schema = MyModel.model_json_schema()
print(json.dumps(schema, indent=2))

Environment

Pydantic: 2

What Broke

Generated JSON schema differs across runs, causing CI validation failures against checked-in schema.

Why It Broke

The JSON schema generation fails to maintain consistent ordering for set-like defaults due to Python's hash randomization

Fix Options (Details)

Option A — Apply the official fix

Ensures deterministic JSON schema defaults by sorting sets, addressing the issue of non-deterministic ordering in JSON schema generation when using unordered collections as defaults.

When NOT to use: This fix is not suitable for cases where the order of set elements is significant.

Fix reference: https://github.com/pydantic/pydantic/pull/12760

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

When NOT to Use This Fix

This fix is not suitable for cases where the order of set elements is significant.

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.

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 generated json schema is non-deterministic when field default is a non empty set (Fix)