Codex/gpt image qwen cliproxyapi - #1557
Conversation
Co-authored-by: Codex <noreply@openai.com>
Co-authored-by: Codex <noreply@openai.com>
|
You do not have enough credits to review this pull request. Please purchase more credits to continue. |
|
Caution Review failedThe pull request is closed. ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Pro Run ID: 📒 Files selected for processing (8)
Disabled knowledge base sources:
📝 WalkthroughSummary by CodeRabbit
WalkthroughThe PR adds dedicated Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~22 minutes Possibly related PRs
Poem
✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Mesa DescriptionTL;DRIntegrates GPT and Qwen image generation capabilities via What changed?File summaries are not available, but the changes likely involve updates to the Description generated by Mesa. Update settings |
![gemini-code-assist[bot]](https://avatars.githubusercontent.com/in/956858?s=60&v=4){kind=small}
There was a problem hiding this comment.
Code Review
This pull request adds support for OpenAI and Qwen providers to the CLI proxy API, updating configuration templates, environment variable handling, and model mappings. It also introduces a mechanism for provider-specific model overrides in the update script. Review feedback suggests centralizing these overrides within the models.json file to avoid maintaining multiple sources of truth and to prevent hardcoded model slugs from becoming stale.
| add_model_override() { | ||
| local key="$1" | ||
| local suffix="$2" | ||
| local value="$3" | ||
| local placeholder="__$(echo "${key}-${suffix}" | tr 'a-z-' 'A-Z_')__" | ||
| sed_args+=(-e "s|${placeholder}|${value}|g") | ||
| } |
There was a problem hiding this comment.
The introduction of add_model_override creates a second source of truth for model versions outside of models.json. This makes maintenance harder as model updates now require changes in both the JSON file and this script.
Consider moving provider-specific model names directly into models.json (e.g., using a key like gpt-image-openrouter). The existing loop in this script will automatically generate the corresponding __GPT_IMAGE_OPENROUTER__ placeholder, allowing you to remove this function entirely.
|
|
||
| # Provider-specific upstream slugs that intentionally differ from the canonical | ||
| # alias stored in models.json. | ||
| add_model_override "gpt-image" "openrouter" "openai/gpt-5.4-image-2" |
There was a problem hiding this comment.
Hardcoding the upstream model slug openai/gpt-5.4-image-2 here bypasses the centralized model management provided by models.json. If the model version is updated in models.json, this hardcoded string will likely become stale or inconsistent.
1 participant
Summary by cubic
Adds image generation via
gpt-image-2and a newqwenprovider tocliproxyapi. Also introduces provider-specific model overrides inllm-update.sh.New Features
openaiprovider withgpt-image-2; mapgpt-image→gpt-image-2; add OpenRouter upstream slugopenai/gpt-5.4-image-2.qwenprovider (DashScope compatible) withqwen3.6-plus.start.sh, and README to supportOPENAI_API_KEYandQWEN_API_KEY(fallback toDASHSCOPE_API_KEY).add_model_overrideinllm-update.shand tests to replace provider-specific placeholders.Migration
OPENAI_API_KEYandQWEN_API_KEY(orDASHSCOPE_API_KEY) in your environment.scripts/llm-update.shto regenerate configs.Written for commit 3968433. Summary will update on new commits.