Nex AGI: Nex-N2.5-Pro

Nex AGI: Nex-N2.5-Pro is a text generation model from Nex AGI, listed on api14 under the model id nex-agi/nex-n2.5-pro. It sits in the text category and is described by its provider as an agentic model built to turn goals into working, verified outcomes. Its stated core strength is agentic coding within a visual feedback loop, which means it is designed to explore codebases and carry out multi-file changes rather than only answer isolated questions.

The model accepts text and image inputs and returns text. Its context window is 262,144 tokens, and the maximum output is 235,929 tokens, so long inputs and long generations are both possible within the same request. On the request side it supports include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, structured_outputs, temperature, tool_choice, tools, top_k, top_logprobs and top_p. The presence of reasoning and reasoning_effort controls, together with tools and tool_choice, points to a model intended for multi-step work where the caller can steer how much internal reasoning happens and let the model call external functions.

Those capabilities suggest a few practical uses. Agentic coding is the clearest fit: pointing the model at a repository, letting it inspect files and dependencies, and having it implement changes across several files before verifying the result. Because it accepts images as well as text, it can also work from screenshots, diagrams or rendered UI states, which suits visual feedback loops where the model checks what a change actually produced. Structured outputs and tool calling make it usable for pipelines that need machine-readable responses or that must query external systems, and the large context window helps when a task requires holding a whole codebase, a long document set or a lengthy conversation in view.

Calling it through api14 works the same way as any other OpenAI-compatible model. You send a request to the chat completions endpoint with the model field set to nex-agi/nex-n2.5-pro, and you can pass any of the supported parameters in the same request body: temperature, top_p, top_k, max_tokens, logprobs and top_logprobs for sampling and token-level inspection, reasoning, reasoning_effort and include_reasoning to control reasoning behaviour, and tools, tool_choice and structured_outputs when you need function calling or a fixed response shape.

Worth noting when choosing between models: output is text only, so anything requiring generated images or audio needs a different model. The very large context and output limits are useful but not always necessary, and for short, simple text tasks a smaller model may be cheaper and faster to work with.