Nace.AI: Drex Max

Nace.AI: Drex Max is a decision model from Nace.AI, and it is the largest model in the provider's Drex line. It is positioned as the successor to Drex v2.0, so teams already using that model can treat Drex Max as the next step up within the same family. Its main category is decisions, which sets it apart from general purpose chat and generation models: the model is built around answering structured questions rather than producing free form prose.

The model accepts text input and returns decisions as output. Its context window is 262,144 tokens, and the maximum output is 235,929 tokens. According to the provider description, Drex Max takes a state, supplied as text or JSON, together with typed questions. The supported question types are yes/no, multiple choice, and ordinal score. That combination of a large context window and typed question formats means you can pass in a substantial body of material as the state and then ask for answers in a shape your application can consume directly, rather than parsing free text.

Practical use cases follow from that design. A pipeline can hand Drex Max a JSON state describing a customer, an order, or a system configuration, and ask yes/no questions such as whether a case should be escalated. Multiple choice questions suit routing and classification, for example picking a category or a next action from a fixed set. Ordinal score questions fit ranking and grading tasks, such as assigning a risk level or a priority score. Because the state can be text as well as JSON, the same model can work over documents, logs, or structured records, and the long context allows large inputs to be considered in a single call.

You call the model through api14's OpenAI-compatible API using the model id "nace-ai/drex-max". The model lists no supported parameters, so requests should rely on the standard fields for the endpoint and avoid assuming extra tuning options are available. The provider release date is 2026-10-21.

Because the output modality is decisions rather than text, Drex Max is not the right choice when you need generated prose, summaries, or conversational replies. It fits best when the task can be expressed as typed questions over a supplied state. If your workload is open ended generation, a text output model will serve you better; if it is structured decision making over large inputs, Drex Max is designed for exactly that.