Iris-3B vs Recraft API

Iris-3B

Code and weights are Apache-2.0, and one Python package runs text-to-image, a 4-step Turbo, depth and 4x restoration from a documented CLI. It is ten days old, with no PyPI package, version tags, changelog or CI, needs an NVIDIA GPU and about 12 GB of weights, and has no server or API of its own.

Recraft API

Native SVG vector generation, not traced raster. No changelog and no error reference.

Iris-3B

1. Check for an NVIDIA GPU with CUDA and Python 3.11 or newer before downloading. The base weights are about 12 GB, and Qwen3-VL-4B downloads on the first run
2. Exclude depth/*, upscaler/* and turbo/* in hf download unless those models are needed. Each of the first two adds about 12 GB
3. Don't pass --steps, --cfg-scale, --negative-prompt, --order or --shift with a Turbo checkpoint. The script exits with an error
4. Pass --seed to reproduce an image, and set --outdir, since files are named after the prompt and a repeat run overwrites them
5. Load only exported directories (model.safetensors and config.yaml). A training .pth is read with torch.load(weights_only=False), so never load one from an untrusted source

Recraft API

1. Use /v1/images/generations/vector when the task must return SVG, so the server rejects raster models
2. Ask for b64_json or multipart if you can't fetch the URL within a day
3. Name recraftv4_1_flash for drafts at $0.007. The default recraftv4_1 costs $0.035
4. Stay under 100 images a minute and 5 requests a second per user
5. Point the OpenAI SDK at https://external.api.recraft.ai/v1, but expect some parameters to be ignored