OpenAI: GPT-5.6 Terra Pro (batch)

GPT-5.6 Terra Pro (batch) is a text model from OpenAI, identified on api14 as openai/gpt-5.6-terra-pro:batch. It shares its underlying model with GPT-5.6 Terra, but it is served with reasoning.mode set to pro, which is aimed at higher-quality responses on complex tasks. The provider released it on 2026-07-09, and its knowledge cutoff is 2026-02-16. The batch label indicates how the model is served through the gateway rather than a different base model.

The model accepts text, image and file inputs, and it returns text only. Its context window is 1,050,000 tokens, which is large enough to hold very long documents, codebases or multi-file inputs in a single request. The maximum output is 128,000 tokens, so long generations are possible within that context. Because pro mode is used, the model spends more compute on reasoning than the standard variant, which is the trade-off behind the higher-quality responses it targets.

Those capabilities point to work where the input is large and the reasoning is hard. Long document analysis, review of contracts or specifications, and question answering over a whole repository fit the combination of a very large context window and file and image input. Structured extraction and classification are also reasonable, since the model supports structured outputs and response_format. Tool use is supported as well, so agent-style workflows that call functions and act on the results are within scope. Image input allows tasks that mix screenshots, diagrams or scanned pages with text instructions.

You call the model through api14's OpenAI-compatible API using the model id openai/gpt-5.6-terra-pro:batch. The supported request parameters are include_reasoning, max_tokens, reasoning, reasoning_effort, response_format, seed, structured_outputs, tool_choice, tools and verbosity. In practice this means you can control the reasoning behaviour and how much of it is returned, cap the output length with max_tokens, request JSON or schema-constrained output, set a seed for reproducibility, and define tools with a tool_choice. The endpoint follows the OpenAI chat completions shape, so existing OpenAI client libraries work with only the base URL and model id changed.

The main consideration is that pro mode spends far more compute than the standard GPT-5.6 Terra, so it suits tasks where response quality matters more than speed or throughput. For simpler prompts, high-volume classification or latency-sensitive interactive use, the standard GPT-5.6 Terra variant is likely the better fit. The batch serving mode also implies that requests are handled asynchronously rather than in a live conversational loop, so it is a poor match for real-time chat. If your task does not need the largest context or the strongest reasoning, a smaller or non-pro model will usually be the more practical choice.