Nex AGI: Nex-N2.5-Mini

Nex AGI: Nex-N2.5-Mini is a text generation model from Nex AGI, listed on api14 under the model id nex-agi/nex-n2.5-mini. It belongs to the Nex-N2.5 family, which Nex AGI describes as agentic models built to turn goals into working, verified outcomes. The mini variant sits within that family, so it inherits the same agentic orientation while being the smaller option in the line. It is a text model in api14's catalogue, and its provider release date is 2026-09-08.

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 it can take in very large inputs and produce long responses in a single call. Nex AGI highlights agentic coding within a visual feedback loop as a core strength, which lines up with the image input support: the model can explore codebases, implement multi-file changes, and work against visual feedback rather than text alone. The supported parameters include reasoning controls (reasoning, reasoning_effort, include_reasoning), sampling controls (temperature, top_p, top_k), output length (max_tokens), structured_outputs, and log probability options (logprobs, top_logprobs).

Those capabilities point to a few practical uses. Agentic coding tasks are the clearest fit: working through an existing repository, making coordinated edits across several files, and iterating on the result. Because images can be supplied as input, tasks where a screenshot, diagram, or rendered interface matters alongside code are also reasonable candidates. The large context window suits work that involves long files, large diffs, or extended conversation history that should stay in view. Structured outputs make it usable for pipelines that need machine-readable responses rather than free text, and the reasoning parameters let you tune how much internal reasoning effort the model applies before answering.

You call it through api14's OpenAI-compatible API using the model id nex-agi/nex-n2.5-mini, so existing OpenAI client libraries and request shapes work without changes beyond the model name and base URL. The request parameters it supports are include_reasoning, logprobs, max_tokens, reasoning, reasoning_effort, structured_outputs, temperature, top_k, top_logprobs, and top_p. Note that top_k is not part of the standard OpenAI parameter set, so you may need to pass it explicitly depending on your client.

On limits: the model outputs text only, so it is not the right choice if you need image, audio, or other non-text output. If your task does not involve images or agentic coding, a plain text model may be simpler. The facts here do not cover pricing, throughput, or quality comparisons, so check the live price table and test against your own workload before committing.