OpenAI lists GPT-6.1 Terra’s standard API rates at $1 per million uncached input tokens, $0.05 per million cached input tokens, and $5 per million output tokens. Those are token-billing rates for API use, not a per-task price. For a request with more than 272,000 input tokens, the higher long-context rates apply to the entire request. OpenAI’s GPT-6.1 Terra model reference lists the rates and billing conditions.
GPT-6.1 Terra’s API rates
All rates in the table are in U.S. dollars per million tokens. The long-context rates apply when a request exceeds 272,000 input tokens; Batch and Flex are half the standard rates, while Fast mode is twice the standard rates.
| Processing and condition | Input per 1 million tokens | Cached input per 1 million tokens | Output per 1 million tokens |
| Standard, up to 272,000 input tokens | $1.00 | $0.05 | $5.00 |
| Long context, above 272,000 input tokens | $2.00 | $0.10 | $7.50 |
| Batch or Flex, short context | $0.50 | $0.025 | $2.50 |
| Fast mode, short context | $2.00 | $0.10 | $10.00 |
Regional processing adds a 10% premium where available. The published Batch, Flex, and Fast rates shown here are short-context rates.
What changes above 272,000 input tokens?
Crossing the threshold raises the rates for the whole request, not just the tokens beyond 272,000. For example, with 270,000 input tokens and 20,000 output tokens, the standard-rate calculation is $0.27 for input plus $0.10 for output, or $0.37. At 300,000 input tokens and the same 20,000 output tokens, the long-context rates apply to the whole request: $0.60 for input plus $0.15 for output, or $0.75. These examples exclude caching and tool charges.
GPT-6.1 Terra’s listed maximum input is 922,000 tokens, within a 1,050,000-token context window. A request can therefore exceed the pricing threshold while remaining below the model’s stated maximum input.
Cached input and cache writes are separate charges
Cached input is billed at $0.05 per million tokens at the standard short-context rate. Writing tokens to the cache has separate listed rates: $1.25 per million tokens for 30-minute retention or $2.50 per million tokens for 12-hour retention, before long-context adjustments. A low cached-input rate does not mean cache writes are free.
Estimate a bill from token use
For an illustration, assume 1,000 calls, each with 10,000 uncached input tokens and 4,000 output tokens. That totals 10 million input tokens and 4 million output tokens. At standard rates, the calculation is (10 × $1) + (4 × $5) = $30 in token charges. This assumes each call stays below the long-context threshold and excludes cache writes, tools, taxes, and any applicable regional premium.
The final API charge depends on how many input and output tokens the requests use, whether input is cached, and which processing mode applies. A rate per million tokens alone does not give a fixed price for a completed task.
API billing and plan access
The API rate card prices usage by tokens. Separately, OpenAI’s October 26, 2026 announcement described GPT-6.1 Terra rolling out to Plus, Pro, Enterprise, and Edu users in ChatGPT Work and Codex. That plan access is a different route from token-billed API use.
