Fix Pydantic AI validation error
Quick answer
A
pydantic_ai validation error occurs when the AI response does not match the defined response_model schema. Fix it by ensuring your Pydantic model fields exactly match the expected AI output and use response_model parameter correctly in client.chat.completions.create. ERROR TYPE
code_error ⚡ QUICK FIX
Ensure your Pydantic model fields match the AI response and pass it as
response_model in the chat.completions.create call.Why this happens
This error arises when the AI's JSON response structure does not conform to the Pydantic model you defined. For example, if your model expects a field answer but the AI returns a different key or nested structure, Pydantic raises a validation error. This typically happens when the response_model schema is mismatched or incomplete.
Example broken code:
import os
from openai import OpenAI
from pydantic import BaseModel
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
class QA(BaseModel):
answer: str
response = client.chat.completions.create(
model="gpt-4o-mini",
response_model=QA,
messages=[{"role": "user", "content": "What is RAG?"}]
)
print(response.answer) output
pydantic.ValidationError: 1 validation error for QA answer field required (type=value_error.missing)
The fix
Adjust your Pydantic model to match the AI response structure. The AI returns the content inside choices[0].message.content, so your model should reflect that. Use response_model to parse the nested content correctly by defining a model that matches the expected JSON shape or by using a custom root type.
Corrected code example:
import os
from openai import OpenAI
from pydantic import BaseModel
client = OpenAI(api_key=os.environ["OPENAI_API_KEY"])
class QA(BaseModel):
answer: str
# Use a wrapper model to parse the AI response content
class AIResponse(BaseModel):
answer: str
response = client.chat.completions.create(
model="gpt-4o-mini",
response_model=AIResponse,
messages=[{"role": "user", "content": "Extract the answer: RAG stands for Retrieval-Augmented Generation."}]
)
print(response.answer) output
Retrieval-Augmented Generation
Preventing it in production
To avoid validation errors in production, always:
- Define Pydantic models that exactly match the expected AI output structure.
- Test your models with sample AI responses before deployment.
- Implement error handling and fallback logic for validation failures.
- Use retries or alternative parsing if the AI output format changes unexpectedly.
Key Takeaways
- Always align your Pydantic
response_modelwith the AI response JSON structure. - Use
response_modelparameter inclient.chat.completions.createto enable automatic validation and parsing. - Test your Pydantic models with sample AI outputs to catch validation issues early.