October 27, 2026 · AI Radar
On the previous Terra, one million output tokens cost $12. On GPT-6.1 Terra, which landed in OpenAI's API catalog last night, it's $5. The input side also dropped from $2 to $1. The new version of the same tier, at roughly half the price of the previous one.
A token is a chunk of the text a model reads and writes; a Turkish word averages several tokens. For a business that has AI answer customer emails, draft proposals or write product descriptions, the bill comes straight out of these two numbers. That's why the price line matters more in this news than the model name.
In OpenAI's model naming, the number indicates the generation, while Sol/Terra/Luna indicate permanent capability tiers. 6.1 Terra is the new version of the middle tier. OpenAI's developer page describes it as a "fast, efficient model for visual reasoning, precise computer use, and clear, natural writing" [A]. The context window is 1,050,000 tokens, it can write up to 128,000 tokens in a single response, and its knowledge cutoff is June 1, 2026. It takes text and images as input; no audio or video, and output is text only [A].
Reasoning effort has six levels: none, low, medium (default), high, xhigh and max [A]. The higher the level, the more the model "thinks," and that thinking is billed as output tokens; so the $5 line can effectively cost more at the max level. How much is not clear in the official source.
One more measure: in benchmarks reported as OpenAI's own tests, Terra is said to reach the same level as GPT-6 Astra at a tenth of the cost [C]. These figures aren't clear on the official page; they come from a third-party summary, so I wouldn't make them the sole basis for a decision.
| Model | Input | Cached input | Output |
|---|---|---|---|
| GPT-6 Astra | 10.00 | 1.00 | 50.00 |
| GPT-6.1 Sol | 2.00 | 0.10 | 10.00 |
| GPT-5.6 Terra (old) | 2.00 | 0.20 | 12.00 |
| GPT-6.1 Terra | 1.00 | 0.05 | 5.00 |
| GPT-6 Luna | 0.10 | 0.01 | 0.50 |
Source: OpenAI API pricing page [A].
Let me run a concrete calculation; this calculation is mine, not OpenAI's. Picture 500 customer emails; each one has about 1,000 tokens read and a 300-token reply written. That comes to 0.5 million input and 0.15 million output tokens in total. On GPT-6.1 Terra, that's $1.25. On the old Terra it's $2.80, on Sol $2.50 and on Astra $12.50. The numbers are low; the real difference shows when there are tens of thousands of requests a month. Thinking tokens aren't included in this calculation.
| Tool | What changed | Label | Who it helps |
|---|---|---|---|
| ChatGPT(in Turkish) | GPT-6.1 Terra opened up in the Work section for Plus, Pro, Business, Enterprise and Edu; not on the Chat screen | New model | Businesses using Work |
| OpenAI Codex CLI(in Turkish) | For Terra, the local message limit in a 5-hour window is 35-285; for Sol, 15-160 (Plus and standard Business) | Price/limit | Software teams |
| Gemini(in Turkish) | Gemini 4 Argon at $2/$10 until November 30; $4/$20 from December 1. Not on the free tier | Price/limit | Anyone using Gemini via the API |
| Gemini(in Turkish) | Gemini 3.8 and 3.6 Flash at $0.75/$3.75; $1.50/$7.50 from January 1, 2027 | Price/limit | Anyone using the cheap Flash model |
| Kimi(in Turkish) | Kimi K3.1 Pro appeared in the OpenRouter catalog; price and license aren't clear in the official source | New model | Anyone looking for an alternative model |
| \n |
No. According to OpenAI's Codex pricing page, it's available in Work and Codex on the Plus, Pro, Business, Enterprise and Edu plans; it can't be used on the Chat screen [A]. On enterprise plans, access depends on workspace settings and the admin.
The page has no note on when the old model will be retired [A], so there's no rush. But when a model doing the same job costs half as much, it deserves a small test.
Because Google's page says it will double on December 1 [A]. If you're using the API today at $2/$10, now is the time to work out your December budget.
If you want to get clear on how to fit an AI model into your business, see AI consulting; if you want to automate your repetitive work, see the AI automations page. To see where you stand first: START YOUR FREE ANALYSIS .
AI Guide
ChatGPTAnswering questions, drafting emails and other text, summarizing documents, writing product descriptions and social media copy; limited deep research on the free plan.(in Turkish)
OpenAI Codex CLIAn open-source coding agent that runs in the terminal. Used with a ChatGPT account (the CLI on Plus and higher plans; sources conflict on whether Free and Go include the CLI) or an API key.(in Turkish)
GeminiChat, writing, image generation and sourced research reports with Deep Research; use inside Gmail and Docs on paid plans.(in Turkish)
Kimi (Moonshot AI)Moonshot AI's chat assistant and Kimi models: it does research and builds documents, presentations and websites; in Agent Swarm mode it splits the work among subagents running in parallel. Developers connect the Kimi API to coding agents; the model weights are open.(in Turkish)Beyond GPT-6, ChatGPT got five new features in October 2026. Audio uploads are on paid plans, Slides on Pro and business plans, locked chats for everyone; Finances is US-only.
+2Google has deprecated Gemini 2.5 Pro and Flash; starting December 1, Gemini API requests are quietly redirected to 3.8 Flash and 3.5 Flash-Lite. What happens to price and quality, and which automations are affected?
+2Anthropic released Claude Haiku 5.5 on October 7: $0.10/$0.50 per million tokens, an effort setting for the first time, and it's selectable on the claude.ai free plan. Which jobs is it good enough for, what do 10,000 customer messages cost, and what does the official source say about Turkish?
+2We use cookies to improve the site and measure advertising performance. Analytics and advertising cookies run only if you accept them. Details are in our Privacy Policy and Data Protection Notice.
