GPT Image 2.5 is live — OpenAI's newest image model, targeted edits that leave the rest of the frame alone
All resources

· reAPI Team

GPT-6.1 Terra Pricing: Token Rates, Cache Costs and Calculator

GPT-6.1 Terra pricing from OpenAI's price page: input, output, cache reads and writes, the 272K long-context tier, Batch and Fast, plus worked cost examples.

GPT-6.1 Terra Pricing: Token Rates, Cache Costs and Calculator

GPT-6.1 Terra pricing starts at $1.00 per million input tokens and $5.00 per million output tokens on OpenAI's Standard tier[1]. OpenAI released the model on October 26, 2026 and describes it as "a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing"[2][3]. The two headline rates are only part of a real bill: cached input, cache writes, a separate tier for prompts over 272K tokens and reasoning tokens all change what a request costs.

This guide lists every GPT-6.1 Terra rate from OpenAI's pricing page, retrieved October 27, 2026, shows how one request is priced, and works through five cost examples with the arithmetic written out, so you can redo them with your own token counts.

TL;DR

  • Standard rates per 1M tokens: $1.00 input, $0.05 cached input, $1.25 cache writes (30-minute lifetime), $5.00 output[1].
  • Over 272K input tokens, the full request moves to $2.00 input, $0.10 cached input, $2.50 cache writes and $7.50 output[1][2].
  • Batch and Flex cost half of Standard; Fast mode costs twice; regional processing adds 10%[2].
  • Reasoning tokens bill as output, so reasoning_effort is the biggest lever after volume[4].
  • Against its siblings: GPT-6.1 Terra costs half of GPT-6.1 Sol on every Standard token type, and ten times GPT-6 Luna for input and output (five times for cached input)[1].
  • On reAPI (checked October 27, 2026): $0.75 input and $3.75 output per 1M tokens, billed per token from one credit balance[5].

GPT-6.1 Terra pricing for every token type

OpenAI prices GPT-6.1 Terra by token type and by prompt length. Short context means up to 272K input tokens; above that, the long-context rates apply to the whole request[1].

Per 1M tokens (Standard)Up to 272K inputOver 272K input
Input$1.00$2.00
Cached input$0.05$0.10
Cache write, 30-minute lifetime$1.25$2.50
Cache write, 12-hour lifetime$2.50$5.00
Output$5.00$7.50

Source: OpenAI API pricing, Standard tier[1].

The model page states the rules behind these numbers. Cached input tokens are "priced at 5% of the uncached input token rate", and "prompts with more than 272K input tokens are priced at 2x input and cache rates and 1.5x output for the full request"[2]. The 5% cache-read ratio is lower than most GPT models: OpenAI's caching guide puts reads at 0.1× input on most recent models and 0.05× only on GPT-6.1 Sol and GPT-6.1 Terra[6].

Batch, Flex and Fast for GPT-6.1 Terra

Standard is not the only price. OpenAI's model page lists three adjustments: "Batch and Flex prices are 50% lower than Standard", "Regional processing adds a 10% premium where available", and "Fast mode prices are 2x Standard"[2].

Per 1M tokens, up to 272K inputInputCached input30-min writeOutput
Batch / Flex$0.50$0.025$0.625$2.50
Standard$1.00$0.05$1.25$5.00
Fast$2.00$0.10$2.50$10.00

Source: OpenAI API pricing, Batch, Flex and Fast tables[1].

One detail stands out: GPT-6.1 Terra at Standard costs exactly what GPT-6.1 Sol costs at Batch, $1.00 input and $5.00 output[1]. If your workload can wait for a batch window, that is the point where the two models cost the same.

How a GPT-6.1 Terra request is priced

Every response reports token counts in usage. The bill for one Standard request below 272K input tokens is:

cost = (uncached_input × $1.00
      + cached_input × $0.05
      + cache_write × $1.25
      + output × $5.00) / 1,000,000

Three things move the result.

Cache. Tokens served from the prompt cache bill at the cache-read rate instead of the input rate. Cache writes cost 1.25× ordinary input for the default 30-minute lifetime, and a 12-hour lifetime chosen in the Responses API costs 2.5×[6]. A long system prompt or reference document that stays identical across calls is where caching pays.

Reasoning. OpenAI's reasoning guide says reasoning tokens occupy the context window "and are billed as output tokens"[4]. GPT-6.1 Terra supports none, low, medium (default), high, xhigh and max[2]. In reAPI's tests on October 27, 2026, a short multi-step arithmetic question used no reasoning tokens at none, 80 at xhigh and 199 at max[7]. Those counts are from one prompt, not a benchmark, but they show the direction.

Prompt length. Cross 272K input tokens and every token type in the request moves to the long-context column, not only the tokens above the line[2].

GPT-6.1 Terra pricing calculator: five worked examples

All five use OpenAI's Standard rates from the tables above. The token counts are illustrative assumptions, chosen to show how each lever moves the total.

ScenarioTokensArithmeticCost
A. Short rewrite2,000 in, 500 out2,000 × $1 + 500 × $5$0.0045
B. Same prefix, cached20,000 in (18,000 cached), 800 out2,000 × $1 + 18,000 × $0.05 + 800 × $5$0.0069
C. Same as B, no cache20,000 in, 800 out20,000 × $1 + 800 × $5$0.024
D. Reasoning at high effort1,000 in, 300 visible + 2,000 reasoning out1,000 × $1 + 2,300 × $5$0.0125
E. Long contract review300,000 in, 4,000 out300,000 × $2 + 4,000 × $7.50$0.63

Each product is per 1M tokens, so divide by 1,000,000. Scenario A at a million requests a day is $4,500. Scenario B is less than a third of C, because the cached share bills at 5% of the input rate. In scenario D, reasoning tokens are most of the bill even though the visible answer is 300 tokens. Scenario E shows the long-context switch: the same 304,000 tokens priced at short-context rates would have cost $0.32, so crossing 272K roughly doubles the request.

How GPT-6.1 Terra pricing compares with other GPT models

Model (Standard, per 1M)InputCached inputOutput
GPT-6 Luna$0.10$0.01$0.50
GPT-6.1 Terra$1.00$0.05$5.00
GPT-6.1 Sol$2.00$0.10$10.00
GPT-5.6 Terra$2.00$0.20$12.00
GPT-6 Astra$10.00$1.00$50.00

Source: OpenAI API pricing, Standard tier[1].

Against GPT-6.1 Sol, every rate is exactly half. Against GPT-6 Luna, every rate is ten times higher, apart from cached input, which is five times. GPT-5.6 Terra, the previous Terra, cost twice as much for input and 2.4 times as much for output[1]. The price gap does not say which model answers better for your task; OpenAI positions Sol for "complex coding, computer use, and professional work" and Terra for visual reasoning and writing[2][8]. GPT-6.1 Terra vs 6.1 Sol covers that choice.

Paying for GPT-6.1 Terra through reAPI

reAPI serves GPT-6.1 Terra on its Chat Completions endpoint with the model id gpt-6.1-terra, billed per token from your reAPI credit balance[7]. On October 27, 2026 the reAPI rates were:

Per 1M tokens on reAPIUp to 272K uncached inputOver 272K uncached input
Input$0.75$1.50
Cached input$0.0375$0.075
Cache write$0.9375$1.875
Output$3.75$5.625

Source: reAPI model page, checked October 27, 2026[5].

Every cell is 75% of OpenAI's Standard rate. One rule differs from OpenAI's: on reAPI the 272K line is measured on uncached input tokens, so cache reads and cache writes do not push a request into the long-context column. Scenario A above costs $0.003375 at these rates, and scenario E $0.4725. The GPT-6.1 Terra model page always shows the current rates and has a playground that sends billed requests. Each response's usage object reports prompt_tokens, completion_tokens (reasoning included) and cached_tokens, so you can reconcile a bill with the formula above[7].

FAQ

gpt terra cache token cost

For GPT-6.1 Terra, cached input costs $0.05 per 1M tokens, 5% of the $1.00 input rate. Cache writes cost $1.25 per 1M for the default 30-minute lifetime and $2.50 for a 12-hour lifetime; above 272K input tokens all three double[1][6]. On reAPI the cache-read rate was $0.0375 per 1M on October 27, 2026[5].

is gpt 6 terra free

There is no model called GPT-6 Terra; OpenAI's Terra in this generation is GPT-6.1 Terra, model id gpt-6.1-terra[2]. Through the API it is billed per token, and OpenAI's pricing page lists only paid rates for it[1].

gpt 6.1 terra pricing calculator

Multiply each token count in usage by its rate and divide by 1,000,000: uncached input × $1.00, cached input × $0.05, cache writes × $1.25 and output × $5.00 at Standard, or the long-context column when the prompt is over 272K tokens[1]. The five examples above show the arithmetic.

gpt 6.1 terra vs 6.1 sol price

GPT-6.1 Terra is half the price of GPT-6.1 Sol on every Standard token type: $1.00 vs $2.00 input, $0.05 vs $0.10 cached input and $5.00 vs $10.00 output per 1M tokens[1]. GPT-6.1 Sol also offers an Ultrafast tier at 6x Standard, which the pricing page does not list for Terra[1][8].

gpt terra vs luna

GPT-6.1 Terra costs ten times GPT-6 Luna for input and output: $1.00 vs $0.10 input and $5.00 vs $0.50 output per 1M tokens[1]. OpenAI calls Luna its "most efficient model for focused, high-volume tasks"; Terra is aimed at visual reasoning, computer use and writing[2][9].

gpt terra cost

A short request with 2,000 input and 500 output tokens costs $0.0045 at OpenAI's Standard rates and $0.003375 at reAPI's rates of October 27, 2026[1][5]. Reasoning effort and prompts over 272K tokens are what push a single request well above that.

Keeping GPT-6.1 Terra costs predictable

Most of a GPT-6.1 Terra bill comes from two places: reasoning tokens and uncached input. Start at medium, which OpenAI recommends, and step down for routine rewrites or tagging where a lower effort holds your accuracy[10]. Keep a stable prefix at the start of every conversation so repeat calls read from cache at 5% of the input rate, and split documents that would push a prompt past 272K tokens. Batch work that can wait drops GPT-6.1 Terra pricing by half again. For current per-token rates and a playground, open the GPT-6.1 Terra model page.

References

  1. OpenAI. API pricing. Retrieved October 2026 from developers.openai.com/api/docs/pricing
  2. OpenAI. GPT-6.1 Terra model page. Retrieved October 2026 from developers.openai.com/api/docs/models/gpt-6.1-terra
  3. OpenAI. API changelog, October 26, 2026. Retrieved October 2026 from developers.openai.com/api/docs/changelog
  4. OpenAI. Reasoning models. Retrieved October 2026 from developers.openai.com/api/docs/guides/reasoning
  5. reAPI. GPT-6.1 Terra model page. Retrieved October 2026 from reapi.ai/models/gpt-6-1-terra
  6. OpenAI. Prompt caching. Retrieved October 2026 from developers.openai.com/api/docs/guides/prompt-caching
  7. reAPI. GPT-6.1 Terra API docs. Retrieved October 2026 from reapi.ai/docs/gpt-6-1-terra
  8. OpenAI. GPT-6.1 Sol model page. Retrieved October 2026 from developers.openai.com/api/docs/models/gpt-6.1-sol
  9. OpenAI. GPT-6 Luna model page. Retrieved October 2026 from developers.openai.com/api/docs/models/gpt-6-luna
  10. OpenAI. GPT-6 guide: GPT-6.1 Terra. Retrieved October 2026 from developers.openai.com/api/docs/guides/latest-model
All resources