DALL-E 3 has been deprecated. some of our examples, SDK, etc use this model so the API call fails #1510

OpenAI officially deprecated and retired DALL·E 3 (along with DALL·E 2)

They now recommend using :

GPT Image 2: This is the current flagship alternative for high-quality, complex creative generation and detailed prompt adherence.
gpt-image-1-mini: This is the fast, cost-efficient alternative designed for high-volume applications or rapid prototyping.

I saw this issue on 2 examples
a. examples/agents/40_media_generation_agent.py
b. examples/agents/kitchen_sink.py

According to a claude code sweep, it says these are all the places we usei it
python-sdk (12)

Examples users copy:
examples/agents/40_media_generation_agent.py:31
examples/agents/kitchen_sink.py:625

SDK source + docstrings:
src/conductor/ai/agents/tool.py:465, 478
src/conductor/client/workflow/task/llm_tasks/generate_image.py:17

Tests and e2e:
tests/integration/test_ai_task_types.py:282
tests/integration/test_ai_examples.py:42 OPENAI_IMAGE_MODEL = "dall-e-3"
e2e/test_suite1_basic_validation.py:468
e2e/test_suite7_media_tools.py:4, 129, 138

java-sdk (10)

docs/agents/concepts/tools.md:218
examples/ai-generation/ai-image-generation/.../GenerateWorker.java:44
e2e/.../Suite7MediaTools.java:85, 99, 204, 226, 260, 339
conductor-client-ai/src/test/.../SerializerTest.java:534, 558

conductor-oss server (7)

ai/README.md:157, 897, 1440
ai/examples/03-image-generation.json:12
ai/examples/13-image-to-video-pipeline.json:13
docs/documentation/clientsdks/ruby-sdk.md:306
docs/devguide/cookbook/ai-llm.md:216