Validation Error for ResponseTextDeltaEvent – Missing logprobs Field #1276
Description
📌 Please read this first
- Have you read the docs? Yes — Agents SDK docs
- Have you searched for related issues? Yes — couldn’t find a matching one for this validation error.
🐛 Describe the bug
When using Runner.run_streamed() and handling the event as ResponseTextDeltaEvent, I encountered a Pydantic validation error due to the logprobs field being missing in the streamed response data.
It seems the model expects logprobs to always be present — even when the API doesn’t include it in the delta. I tried using this setup, but the error consistently occurs.
🧩 Debug information
- Agents SDK version:
v0.2.3 - Python version:
3.12.9 - Platform: Windows 10
- LLM provider: Using Gemini API (OpenAI compatible format)
🔁 Repro steps
Here’s a minimal reproducible example:
from agents import Runner
from openai.types.responses import ResponseTextDeltaEvent
# Example agent and prompt
result = Runner.run_streamed(agent, "What is the capital of France?")
async for event in result.stream_events():
if event.type == "raw_response_event" and isinstance(event.data, ResponseTextDeltaEvent):
print(event.data.delta) # This line triggers the validation error❌ Actual Error
Error: 1 validation error for ResponseTextDeltaEvent
logprobs
Field required [type=missing, input_value={'content_index': 0, 'del...', 'sequence_number': 3}, input_type=dict]
For further information visit https://errors.pydantic.dev/2.11/v/missing
✅ Expected behavior
The SDK should handle streamed delta events gracefully, even when logprobs is not included in the response payload.
This field should be marked as optional (e.g., Optional[...] = None) in the Pydantic model to prevent validation failures for missing data.
🙋 Extra Notes
Let me know if I can assist with debugging or testing a fix — I’m currently studying and deeply exploring the OpenAI Agents SDK.