It would be better if OpenAI would report instead of me re-checking against resources and re-running code..
###status check June 5
- 1a. example python for “check batch status” - docs FIXED
- 1b. purposes of batch output, on files API, only for upload purposes - api ref PARTIAL
-
- “for reasoning models like gpt-4.1…” - help UNFIXED (https://help.openai.com/en/articles/5072518-controlling-the-length-of-openai-model-responses)
-
- “view limits page to check verification model access” (gpt-image-1 shows in model list for unpaid org) - help UNFIXED
-
- Assistants API reference, “create assistant”
tool_resources
->file search - api reference UNFIXED
- Assistants API reference, “create assistant”
-
- API Reference → Responses → Delete model response ID example: API ref FIXED
- 6a. API Reference create image->
"quality":"standard" is the only option for dall-e-2
doesn’t work. "Unknown parameter: ‘quality’ UNFIXED tested
6b. same issue for edits: dall-e-2 won’t accept “standard” for quality as documented.
(another half-dozen remain unchecked)
Here’s code for getting the dall-e-2 error by following documentation:
from PIL import Image; import base64, io
from openai import OpenAI; client = OpenAI()
# Create a blank black transparent RGBA PNG image (dalle creates everything)
image = Image.new("RGBA", (1024, 1024), (0, 0, 0, 0))
image_bytes = io.BytesIO()
image.save(image_bytes, format='PNG')
image_bytes.name = "blank_black.png" # filename required for mimetype detection
image_bytes.seek(0) # Reset file pointer to the beginning
# API Call with image file-like object
resp = client.images.edit(
model="dall-e-2",
prompt="A cute baby sea turtle",
quality="standard", # 'message': "Unknown parameter: 'quality'."
size="1024x1024",
image=image_bytes
)
print(resp.data[0].url)