Spring AI: OpenAI API returns 400 Bad Request with no detailed error message

[hedaoming]

I encountered an issue when using Spring AI to call OpenAI API. The request only returns a 400 Bad Request error without any detailed error information.

Environment:

• Spring AI version: 1.0.0-M5
• JDK version: 21
• Dependencies:

<dependency>
    <groupId>org.springframework.ai</groupId>
    <artifactId>spring-ai-bom</artifactId>
    <version>1.0.0-M5</version>
    <type>pom</type>
    <scope>import</scope>
</dependency>

Error Log:

2025-03-14 10:50:48.251 ERROR 400 --- [reactor-http-epoll-3] [4c890702b250425c82245f422ed21aa1] [] : Aggregation Error
org.springframework.web.reactive.function.client.WebClientResponseException$BadRequest: 400 Bad Request from POST https://api.openai.com/v1/chat/completions
        at org.springframework.web.reactive.function.client.WebClientResponseException.create(WebClientResponseException.java:307)
        Suppressed: The stacktrace has been enhanced by Reactor, refer to additional information below:
Error has been observed at the following site(s):
        *__checkpoint ⇢ 400 BAD_REQUEST from POST https://api.openai.com/v1/chat/completions [DefaultWebClient]
Original Stack Trace:
                at org.springframework.web.reactive.function.client.WebClientResponseException.create(WebClientResponseException.java:307)
                at org.springframework.web.reactive.function.client.DefaultClientResponse.lambda$createException$1(DefaultClientResponse.java:214)
                at reactor.core.publisher.FluxMap$MapSubscriber.onNext(FluxMap.java:106)

When making the same request using curl with identical parameters:

curl https://api.openai.com/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer xxxxx" \
  -d '{
    "model": "o1-mini",
    "messages": [
      {
        "role": "system",
        "content": "You are a helpful assistant."
      },
      {
        "role": "user",
        "content": "Hello!"
      }
    ],
    "stream": true
  }'

The curl request revealed the actual error message from OpenAI:

{
  "error": {
    "message": "Unsupported value: 'messages[0].role' does not support 'system' with this model.",
    "type": "invalid_request_error",
    "param": "messages[0].role",
    "code": "unsupported_value"
  }
}

Question: Why doesn't Spring AI display these detailed error messages from OpenAI? It would be much more helpful for debugging if the actual error response from OpenAI was included in the logs. Is there a way to configure Spring AI to show the complete error response?

Any help or insights would be greatly appreciated!

[dev-jonghoonpark]

Could you please provide the code for the calling part?