pub struct RawLiteLLMEntry {Show 23 fields
pub litellm_provider: Option<String>,
pub mode: Option<String>,
pub max_input_tokens: Option<u32>,
pub max_output_tokens: Option<u32>,
pub max_tokens: Option<u32>,
pub input_cost_per_token: Option<f64>,
pub output_cost_per_token: Option<f64>,
pub cache_read_input_token_cost: Option<f64>,
pub cache_creation_input_token_cost: Option<f64>,
pub output_cost_per_reasoning_token: Option<f64>,
pub input_cost_per_audio_token: Option<f64>,
pub deprecation_date: Option<String>,
pub supports_function_calling: Option<bool>,
pub supports_vision: Option<bool>,
pub supports_audio_input: Option<bool>,
pub supports_audio_output: Option<bool>,
pub supports_video_input: Option<bool>,
pub supports_web_search: Option<bool>,
pub supports_pdf_input: Option<bool>,
pub supports_reasoning: Option<bool>,
pub supports_computer_use: Option<bool>,
pub supports_prompt_caching: Option<bool>,
pub supports_assistant_prefill: Option<bool>,
}Expand description
Raw LiteLLM entry as it appears in the JSON.
Fields§
§litellm_provider: Option<String>Provider slug as LiteLLM reports it.
mode: Option<String>What the model does: chat, embedding, image_generation, and so on.
max_input_tokens: Option<u32>Largest prompt the model accepts, in tokens.
max_output_tokens: Option<u32>Largest completion the model will produce, in tokens.
max_tokens: Option<u32>Combined limit, used when the input/output split is unknown.
input_cost_per_token: Option<f64>USD charged per input token.
output_cost_per_token: Option<f64>USD charged per output token.
cache_read_input_token_cost: Option<f64>USD per input token served from the prompt cache.
cache_creation_input_token_cost: Option<f64>USD per input token written to the prompt cache.
output_cost_per_reasoning_token: Option<f64>USD per reasoning token, where billed separately.
input_cost_per_audio_token: Option<f64>USD per audio input token.
deprecation_date: Option<String>Date the provider retires the model, as YYYY-MM-DD.
supports_function_calling: Option<bool>Whether the model can call tools.
supports_vision: Option<bool>Whether the model accepts images.
supports_audio_input: Option<bool>Whether the model accepts audio.
supports_audio_output: Option<bool>Whether the model can emit audio.
supports_video_input: Option<bool>Whether the model accepts video.
supports_web_search: Option<bool>Whether the provider offers built-in web search.
supports_pdf_input: Option<bool>Whether the model accepts PDFs directly.
supports_reasoning: Option<bool>Whether the model exposes reasoning or thinking output.
supports_computer_use: Option<bool>Whether the model supports computer-use tooling.
supports_prompt_caching: Option<bool>Whether the provider caches prompt prefixes.
supports_assistant_prefill: Option<bool>Whether an assistant turn may be prefilled.