GPT Image 2.5 moderation low: can you send it through Sume?

OpenAI's guide lists a moderation parameter for GPT Image with two values, auto (default) and low. Sume's Image API request table does not list moderation, and a parameter the selected model does not list is rejected with 400 unsupported_parameter.

So do not assume moderation: "low" passes through. Read the model's descriptors first. Pages read 2026-09-30.

What does OpenAI say about moderation?

The guide lists moderation with auto as the default and low as the other value. This post quotes no further meaning for either value.

What does Sume list instead?

Sume's request parameters for POST /v1/images include model, prompt, n, resolution, aspect_ratio, size, quality, output_format, mask_url, background, input_references, provider.*, metadata, mode, webhook_url and wait_timeout_seconds. moderation is not among them. allowed_passthrough_parameters is empty for every endpoint in v1, so provider-specific keys cannot be smuggled through provider.options either.

How do I check what a model accepts?

Call GET /v1/images/models/openai/gpt-image-2.5/endpoints and read supported_parameters. That is the definitive set for the endpoint; if moderation is not in it, sending it will fail.

curl "https://api.sume.com/v1/images/models/openai/gpt-image-2.5/endpoints" \
  -H "Authorization: Bearer $SUME_API_KEY"

What if a prompt is refused?

Rewrite the prompt rather than looking for a filter switch. Sume's docs do not describe a way to loosen content filtering on GPT Image 2.5, so omit any claim that one exists.

How do I check this myself?

Sume's page could add parameters later, so the catalog call is the source of truth, not a blog post. Re-run it when you change models, because each model publishes its own supported_parameters. The linked docs pages and the catalog endpoint show the current values, and this post reflects them as of 2026-09-30.