GPT 5.6 Terra
Neural Network
GPT 5.6 Terra for code and reasoning — a balanced OpenAI model from the GPT-5.6 series for everyday and agentic tasks.
Specifications
- Input
- text, images, documents
- Output
- text
- Context size
- 1 050 000 tokens
- Max answer length
- 128 000 tokens
- Reasoning effort
- none, low, medium, high, extra high, max
- Capabilities
- reasoning, function calling
Prompt cost
(per 1M tokens)
Answer cost
(per 1M tokens)
How it works GPT 5.6 Terra?
GPT 5.6 Terra sits in the middle of OpenAI's GPT-5.6 lineup, between the flagship Sol and the economical Luna, making it the go-to for balancing quality and token usage. It is designed for daily coding, reasoning, and agentic scenarios, with a batch variant for bulk processing. BotHub provides access without VPNs or foreign cards, with pay-as-you-go billing in rubles for non-expiring tokens. Other models are available in the same window for easy comparison, and switching is seamless via a unified OpenAI-compatible API. Dialogues are encrypted, and corporate clients get contracts, invoices, EDI, and an admin panel with limits. It is typically used for code analysis and refactoring, writing tests, explaining modules, building agents with external tools, step-by-step technical and analytical problem solving, and bulk data processing—labeling, classification, summarization—as well as a default workhorse model when heavy requests are left to the series' top tier.Frequently asked questions about GPT 5.6 Terra
You can use generated results for commercial purposes. You own all rights to the content you create. The only restriction: make sure your prompt does not include copyrighted third-party material. You are responsible for respecting the rights to any input data.
This is a text model with a 1,050,000 token context. You can upload documents and images, analyze large sets of files, write and refactor code, find bugs, solve math and STEM problems, prepare analytics, and connect external tools via function calling.
The output limit is up to 128,000 tokens per response. This is enough for a large code module, detailed technical documentation, analysis of a long contract, or a series of related report sections without manually assembling the response from many small parts.
Yes, the reasoning mode is confirmed in our data. The model builds a chain of steps before the final answer, which helps with multi-step tasks: deriving formulas, debugging logic, planning architecture, checking conditions in long documents, and verifying its own intermediate conclusions.
Yes, both function calling and structured JSON output are supported. The model is easy to integrate into agents and pipelines: it calls your tools, accesses databases and APIs, and returns the answer in a pre-defined schema suitable for parsing.
It accepts text, images, and documents as input, including native file processing: diagrams, interface screenshots, tables, scans, and PDFs. Audio and video are not marked as supported input formats in our data. The model generates text as output.
The model belongs to the GPT line; such models are trained on multilingual corpora and work confidently with Russian texts and code. It is best to check the quality with your own materials: in BotHub, you can compare the response with other models in the same window.