MiniMax: MiniMax M2.7

MiniMax M2.7 is a text generation model from MiniMax, listed on api14 under the model id minimax/minimax-m2.7. It sits in the text category and is described by the provider as a next-generation large language model designed for autonomous, real-world productivity and continuous improvement, with agentic capabilities built in through multi-agent work. If you are choosing a general purpose text model for an application, this is the provider's current entry in that line.

The model takes text as input and returns text as output, so it fits standard prompt and completion workflows rather than image or audio tasks. Its context window is 204800 tokens, which is large enough to hold long documents, extended conversations or sizeable code files in a single request. The maximum output is 131072 tokens, so responses can run very long when you need them to. Reasoning output is supported, which means the model can expose its thinking steps alongside the final answer if you ask for them.

Those capabilities point to a few practical uses. Long context plus long output suits document analysis, summarisation of lengthy reports, and drafting work where the model needs to hold a lot of source material and produce a substantial result. Tool calling and reasoning support make it a reasonable fit for agent style workflows, where the model decides which functions to call and works through a task in steps. The multi-agent angle in the provider description suggests it is aimed at exactly this kind of autonomous productivity work rather than single turn chat alone.

You call it through api14's OpenAI-compatible API using the model id minimax/minimax-m2.7, so existing OpenAI client libraries work with a change of base URL and key. The supported request parameters are frequency_penalty, include_reasoning, logit_bias, max_tokens, min_p, presence_penalty, reasoning, repetition_penalty, response_format, seed, stop, temperature, tool_choice, tools, top_k and top_p. That covers sampling controls, determinism through seed, structured output through response_format, and tool use through tools and tool_choice. Use include_reasoning or reasoning when you want the model's reasoning surfaced.

On limits, the facts here do not list restrictions beyond the context and output sizes, so the main constraint to plan around is the 204800 token context and the 131072 token output ceiling. If your task is not text based, or you do not need reasoning or tool calling, a simpler model may be a better fit. For long context, agentic and tool driven text work, this model covers those needs directly.