OpenAI: GPT Terra Latest

GPT Terra Latest is a text model from OpenAI. The provider describes it as a model that always redirects to the latest model in the GPT Terra family, so the underlying version can change over time while the model id stays the same. It sits in the text category, and its knowledge cutoff is 2026-02-16.

The model accepts text, images and files as input, and it produces text as output. Its context window is 1,050,000 tokens, which is large enough to hold very long documents alongside a conversation, and the maximum output is 128,000 tokens. Because images and files are accepted alongside text, you can send mixed prompts that combine written instructions with visual or file based material in a single request.

Those capabilities point to a few practical patterns. Long context work such as summarising or querying large documents, codebases or transcripts fits the 1,050,000 token window well. Image and file input suit tasks like extracting information from screenshots, scanned pages or attached documents, and the 128,000 token output ceiling helps when you need a long structured result rather than a short answer. The reasoning related parameters also make it a reasonable choice for problems where you want the model to work through steps before answering, and the tool and structured output parameters support pipelines that need to call functions or return machine readable responses.

You call it through api14's OpenAI-compatible API using the model id "~openai/gpt-terra-latest". The request parameters it supports are include_reasoning, max_completion_tokens, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, tool_choice, tools and verbosity. In practice this means you can control output length with max_completion_tokens or max_tokens, steer reasoning behaviour with reasoning, reasoning_effort and include_reasoning, request JSON shaped answers with response_format or structured_outputs, define and select functions with tools and tool_choice, and adjust how verbose the reply is with verbosity. The seed parameter is available when you want more repeatable sampling.

One thing to keep in mind is that the model id points at whichever GPT Terra model is currently the latest, so behaviour can shift when OpenAI updates the family, which matters if you depend on highly stable outputs. Output is text only, so anything that needs generated images or audio would require a different model. The knowledge cutoff of 2026-02-16 also means questions about events after that date rely on what you supply in the prompt rather than on the model's own knowledge.