GPT-5.6 Terra

GPT-5.6 Terra is a chat-oriented large language model from OpenAI, listed in our catalog as part of the GPT-5.x line. Our database entry for this model is intentionally sparse: we record the creator, the model type, and independently measured performance figures, and we do not currently track a confirmed context window, modality list, or published benchmark scores for it. Where those details are absent below, that reflects gaps in our data rather than an absence of the capability in the model itself. The performance data we do carry comes from Artificial Analysis, which measured GPT-5.6 Terra at roughly 98.8 output tokens per second with a time to first token of about 2,156 milliseconds. That combination — a healthy sustained generation rate paired with a first-token delay of just over two seconds — is characteristic of models that do some amount of work before the first token appears. In practice it means the model streams responses at a comfortable reading pace once it starts, but interfaces that feel snappy on very short prompts will show a noticeable pause at the beginning of each turn. As a chat model, GPT-5.6 Terra is intended for conversational and instruction-following workloads: assistants, drafting, question answering, summarization, and similar text-in/text-out tasks. Buyers comparing it against other OpenAI entries or third-party alternatives on this site should weigh the measured throughput and latency against provider pricing shown in the table on this page, and consult OpenAI's own documentation for the authoritative specification of context length, supported modalities, and API features.