Mistral: Mistral Large 4

Mistral Large 4 is a text generation model from Mistral, listed on api14 under the model id mistralai/mistral-large-4-0. It sits in the text category and is described by the provider as a frontier multimodal model built for reasoning, coding, and agentic workloads. If you are already working with Mistral models, this is the large, general purpose option in that line.

The model accepts text and image input and returns text. Its context window is 1,048,576 tokens, and the maximum output length is 262,144 tokens. That combination of a very large input window and a large output budget means you can send long documents, codebases, or multi turn histories in a single request and still leave room for a substantial generated response. Image input is supported alongside text, so prompts can mix written instructions with visual material.

The stated focus on reasoning, coding, and agentic workloads maps to a few practical patterns. Long context reasoning suits tasks where the model has to hold a large body of material in view at once, such as reviewing a repository, summarising or querying lengthy documentation, or working through multi step analysis without chunking. Coding tasks benefit from both the context size and the output budget, since generated code and explanations can be long. Agentic use follows from tool support: the model can be given tools and a tool choice policy, which is what you need for workflows where the model decides when to call an external function. Image input extends this to prompts that include screenshots, diagrams, or scanned pages.

You call the model through api14's OpenAI-compatible API using the model id mistralai/mistral-large-4-0, so existing OpenAI client libraries work with a change of base URL and model name. The supported request parameters are frequency_penalty, include_reasoning, logprobs, max_tokens, presence_penalty, reasoning, reasoning_effort, response_format, seed, stop, structured_outputs, temperature, tool_choice, tools, top_logprobs, and top_p. In practice that covers sampling controls (temperature, top_p, frequency and presence penalties, seed, stop), output length (max_tokens), structured responses (response_format and structured_outputs), token level inspection (logprobs and top_logprobs), and tool use (tools and tool_choice). The reasoning related parameters (reasoning, reasoning_effort, include_reasoning) let you control and surface the model's reasoning behaviour where you need it.

There are no stated restrictions in the facts above beyond the modality split: input can be text or image, but output is text only, so this is not a model for generating images. The 1M token context and 262,144 token output ceiling are large, but if your workload is short prompts with tight latency or cost constraints, a smaller model may be a better fit. For long context reasoning, coding, and tool driven agent flows, Mistral Large 4 is the appropriate choice in this lineup.