BREAKING CHANGE: Model is now a string-based struct with constructor
methods instead of a closed enum. EmbedModel is removed — use
Model::mistral_embed() instead. Tool parameters now accept
serde_json::Value (JSON Schema) instead of limited enum types.
- Replace Model enum with flexible Model(String) supporting all
current models: Large 3, Small 4, Medium 3.1, Magistral, Codestral,
Devstral, Pixtral, Voxtral, Ministral, and arbitrary strings
- Remove EmbedModel enum (consolidated into Model)
- Chat: add frequency_penalty, presence_penalty, stop, n, min_tokens,
parallel_tool_calls, reasoning_effort, json_schema response format
- Embeddings: add output_dimension and output_dtype fields
- Tools: accept raw JSON Schema, add tool call IDs and Required choice
- Stream delta content is now Option<String> for tool call chunks
- Add Length, ModelLength, Error finish reason variants
- DRY HTTP transport with shared response handlers
- Add DELETE method support and model get/delete endpoints
- Make model_list fields more lenient with Option/default for API compat
* feat(Model)!: Update model constants
- Add new models Mistral Nemo and Codestral Mamba
- Add aliases to models constant to be deserialized seemlessly from the versioned forme (e.g.: "mistral-large-2407")
- The commonly named Mistral Large 2 is now `constants::Model::MistralLargeLatest` when the old one is `Model::MistralLarge`
* tests(Model): Update the tests to check the new model variants.
BREAKING CHANGE:
- `Chat::ChatParams.safe_prompt` & `Chat::ChatRequest.safe_prompt` are now `bool` instead of `Option<bool>`. Default is `false`.
- `Chat::ChatParams.temperature` & `Chat::ChatRequest.temperature` are now `f32` instead of `Option<f32>`. Default is `0.7`.
- `Chat::ChatParams.top_p` & `Chat::ChatRequest.top_p` are now `f32` instead of `Option<f32>`. Default is `1.0`.
BREAKING CHANGE: You can't set the `stream` option for `client.chat*()`.
Either use `client.chat_stream()` if you want to use streams
or use `client.chat()` / `client.chat_async()` otherwise.
