What model ID should I use for image generation with codex account?

[Bennett-Yang]

I tried gpt-image-2.5 and nanobot told me model_not_found. What's the correct model id? or it's not available for ChatGPT Plus subscription?

[Medamineelkhattabi]

Use gpt-image-2. It's the only image model nanobot's Codex provider is set up for, and gpt-image-2.5 isn't a model ID OpenAI publishes, so the backend rejects it with model_not_found.

I checked v0.4.1. With "provider": "openai_codex", nanobot takes tools.imageGeneration.model, strips an openai-codex/ prefix if there is one, and POSTs it with your ChatGPT OAuth token to https://chatgpt.com/backend-api/codex/images/generations. The error body from that endpoint is passed back to you as-is (image_generation.py#L1358-L1455). nanobot doesn't validate the ID, so whatever you type in the model field goes straight to OpenAI. The provider lists gpt-image-2 as its only option (L1367), the docs say the same (image-generation.md#L121-L139), and OpenAI's own Codex client hardcodes gpt-image-2 for this endpoint (tool.rs#L68).

{
  "tools": {
    "imageGeneration": {
      "enabled": true,
      "provider": "openai_codex",
      "model": "gpt-image-2"
    }
  }
}

Log in with nanobot provider login openai-codex first. The image model only goes in tools.imageGeneration. Your chat model stays a Codex chat model like openai-codex/gpt-5.6-sol.

On Plus: nanobot's docs list Plus or Pro as the requirement for the Codex provider (configuration.md#L745). I couldn't find anything official on per-plan image limits.

If you specifically want Images 2.5, OpenAI publishes it as gpt-image-2.5-flare and gpt-image-2.5-sunburst for the API's /v1/images/generations (model page). I haven't seen anything saying the Codex endpoint accepts those IDs. They're meant for "provider": "openai" with providers.openai.apiKey, which is billed as API usage, not covered by Plus.

[Bennett-Yang]

Thx a lot. It works for me.