Expand description
Auto-generated model catalog with metadata for all known LLM models.
The catalog is generated at build time by the aether-llm-codegen crate from
models.json. Do not edit generated.rs by hand – modify models.json and rebuild instead.
§LlmModel enum
The central type is LlmModel, with one variant per provider family:
Anthropic(AnthropicModel), OpenRouter(OpenRouterModel), Gemini(GeminiModel),
Ollama(String), LlamaCpp(String), etc.
Key methods on LlmModel:
all()– All models in the catalog (static slice).provider()– Provider key (e.g."anthropic","openai").model_id()– The model identifier sent to the API.context_window()– Context window in tokens, if known.provider_display_name()– Human-readable provider name.required_env_var()– The env var needed to use this model.
LlmModel implements FromStr and Display for parsing and formatting
"provider:model_id" strings (e.g. "anthropic:claude-opus-4-6").
§Helper functions
available_models()– Returns catalog models whose required env var is set.get_local_models()– Returnsavailable_models()plus any locally discovered models (Ollama instances, llama.cpp servers).
Enums§
- Anthropic
Model - Bedrock
Foundation Model - Bedrock
Model - Codex
Model - Deep
Seek Model - Gemini
Model - LlmModel
- A model from a specific provider
- Moonshot
Model - Open
Router Model - Openai
Model - Provider
- Typed provider identifier — covers both catalog providers
(
Anthropic,Codex, …) and dynamic providers whose model name is user-supplied (Ollama,LlamaCpp). - ZAiModel
Functions§
- available_
models - Returns models whose provider env var is set
- get_
local_ models - Returns available catalog models plus any locally discovered models.