Grok Imagine Image 2.0 API (`grok-imagine-image-2.0`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Nano Banana Pro (official) API (`gemini-3-pro-image-preview-official`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
FLUX 2 Max API (`flux-2-max`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Z-Image Turbo API (`z-image-turbo`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Z-Image Turbo API (`z-image-turbo`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Z-Image Turbo API (`z-image-turbo`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
FLUX 2 Max API (`flux-2-max`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Nano Banana 2 (official) API (`gemini-3.1-flash-image-preview-official`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Nano Banana (official) API (`gemini-2.5-flash-image-preview-official`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Nano Banana Pro (official) API (`gemini-3-pro-image-preview-official`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
A modern 124M-parameter decoder-only LLM built from scratch. RoPE Β· RMSNorm Β· SwiGLU Β· GQA Β· KV Cache Β· BPE Tokenizer Β· Full training pipeline. Research-grade, production-quality.
GPT Image 2 (official) API (`gpt-image-2-official`): model id, pricing, parameters and copy-paste examples for the OpenAI-compatible APIMart endpoint.
Seedance 2.5 prompt library with reusable video prompts, prompt categories, API quickstarts, parameter guidance, and Argolink examples.
Attribute an image to the model that generated it β from raw pixels, with a ~6M-parameter CNN (RPA).
SHADOW 50M Instruct: a proof of concept. 44M ternary parameters, exact circuits inside the model, memory on disk, 19.8 MB, offline on a CPU.
Validates Goose recipe files for schema errors, missing parameters/defaults, invalid requirements, extension configuration issues, and template mismatches. Understands YAML block scalars and Jinja-style {% for %} / {% set %} locals to minimize false positives.
goose-recipe-graph maps and validates sub_recipe dependencies across a whole recipe library, catching broken references, cycles, and mismatched parameter values. It renders the graph as Mermaid and grades A-F for CI, one level up from recipe-lint which checks a single file.