GPT-6.1 Terra reaches the API, with tool calling reserved for Responses

OpenAI [released GPT‑6.1 Terra on October 26](https://developers.openai.com/api/docs/changelog), listing the model for the Responses and Chat Completions APIs. The [model reference](https://developers.openai.com/api/docs/models/gpt‑6.1‑terra) identifies it as `gpt‑6.1‑terra`, with text and image input and text output. Audio and video are not supported model modalities. OpenAI describes uses such as matching product photographs to catalog entries, interpreting technical drawings, acting on application screens and revising text.

The [reference makes one interface limit explicit](https://developers.openai.com/api/docs/models/gpt‑6.1‑terra): tool calling requires the Responses API. A Chat Completions release listing should therefore not be read as equivalent tool access. Streaming and structured outputs are listed as supported, while fine‑tuning is not. Under Responses, the supported‑tool list includes web search, file search, image generation, code interpreter, hosted shell, apply patch, skills, computer use, MCP and tool search. The base model’s text‑output modality and the separate image‑generation tool are different parts of that interface. A listed tool does not establish that an application has configured the necessary connection, permissions or successful execution.

For `reasoning.effort`, the [documented choices](https://developers.openai.com/api/docs/models/gpt‑6.1‑terra) are `none`, `low`, `medium`, `high`, `xhigh` and `max`; `medium` is the default. The reference gives a context window of 1,050,000 tokens and a maximum output of 128,000 tokens, and marks reasoning‑token support. Its listed knowledge cutoff is June 1, 2026.

The [model page](https://developers.openai.com/api/docs/models/gpt‑6.1‑terra) lists Standard text‑token rates of $1 per million input tokens and $5 per million output tokens. Cached input is $0.05 per million tokens. Cache writes are listed at $1.25 per million tokens for 30‑minute retention and $2.50 for 12‑hour retention, before long‑context adjustments. For prompts above 272,000 input tokens, the page applies 2× input and cache rates and 1.5× output rates to the full request. It also lists separate adjustments for Batch and Flex, regional processing and Fast mode. Tool calls can carry additional charges. The base input/output pair alone therefore does not describe every request’s cost; this report has not made a priced API call.

The [captured Snapshots section](https://developers.openai.com/api/docs/models/gpt‑6.1‑terra) lists `gpt‑6.1‑terra`, without a separate date‑labelled identifier. Use that exact published identifier in any initial integration; the captured page does not establish a distinct dated version that can be pinned. The reference marks the Free tier as unsupported and says rate limits depend on usage tier. It does not establish entitlement for a particular account. OpenTools has reviewed the release note and model reference, but has not sent a request, tested tool calls or verified this model in a third‑party client. *1515 Third Street in San Francisco, home to OpenAI’s headquarters when photographed on June 15, 2025. Archive company context photograph. “1515 Third Street.” Photo: [Coolcaesar](https://commons.wikimedia.org/wiki/User:Coolcaesar), via [Wikimedia Commons](https://commons.wikimedia.org/wiki/File:1515_Third_Street.jpg), [CC BY4.0](https://creativecommons.org/licenses/by/4.0/). Resized and cropped for display.*