Quick Answer
Gemini 4 Argon is the stronger model for hard, long-running work like big coding jobs and deep research. GPT-6.1 Terra is the better everyday pick: faster, half Argon's current API price, and built for computer use and writing. Neither is on a free plan. Terra needs ChatGPT Plus, Argon needs Google AI Pro.
Gemini 4 Argon is the more powerful model, and GPT-6.1 Terra is the better value for most daily work. Argon is Google's new flagship, aimed at long, difficult tasks. Terra is OpenAI's new mid-tier model, aimed at doing ordinary work quickly and cheaply. They aren't direct rivals on paper, but if you're choosing a $20-class plan this month, they're the two new models you're choosing between.
That's worth stating up front, because the headlines make them sound like equals. Google calls Argon its "new frontier model." OpenAI positions Terra below its own flagship, GPT-6 Astra, and still recommends GPT-6.1 Sol or Astra for the most complex coding. So a straight "which is smarter" contest would be unfair to Terra and not very useful to you.
The more practical question is what each one costs, where you can use it, and which kind of work it suits. This comparison sticks to what OpenAI and Google publish on their own pages, plus clearly labelled vendor benchmark claims.
Which Is Better, GPT-6.1 Terra or Gemini 4 Argon?
It depends on the size of the job.
For hard, long-horizon work, Argon. Google built it to "sustain deep reasoning across complex, long-horizon workflows," and raised its output limit to 1M tokens, up from 64K, so it can produce very long code changes or reports in one go. Google's own examples include migrating large C and C++ codebases to Rust and analysing fleet-wide data centre telemetry.
For fast, everyday work, Terra. OpenAI describes it as "a fast, efficient model for visual reasoning, precise computer use, and clear, natural writing." It's the model you'd want reading a screen, filling a form, drafting an email or working through a spreadsheet, again and again, without burning money or time.
If you only remember one thing: Argon for the few tasks that are genuinely hard, Terra for the many that aren't.
What Did OpenAI and Google Release This Month?
Google opened up Argon. Google announced Gemini 4 Argon at the end of September, but only for a group of trusted cyber defenders. This month it became generally available. Google's Gemini app post says "Google AI Pro and Ultra subscribers can now access Gemini 4 Argon," and the Gemini API pricing page now lists it for paid developers.
OpenAI added Terra. OpenAI's developer announcement says GPT-6.1 Terra is rolling out to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and to developers as gpt-6.1-terra. OpenAI calls it a significant step up over GPT-5.6 Terra, and it sits alongside GPT-6 Astra and GPT-6.1 Sol.
Our ChatGPT review and Gemini review have both been updated with these changes.
How Do the Benchmarks Compare?
Not directly, and that's the first thing to know. The two companies reported results on different tests, so there's no shared scoreboard yet. Treat every number below as a vendor claim until independent testers catch up.
Google's numbers for Argon, from its announcement:
- 77.9% on DeepSWE v1.1, a test of real-world, long-horizon software engineering, which Google calls a new state of the art.
- 51.3% on AutomationBench, Zapier's test of end-to-end business tasks, where Google says Argon ranks first.
- 91.7% on LVBench, a long video understanding test.
- 68% on CWE-bench v1, which measures fixing security vulnerabilities, tied for first.
Google also says Argon leads the Vals Index, which weights finance, coding, legal and tax work by their share of US GDP.
OpenAI's numbers for Terra, as reported by BleepingComputer from OpenAI's launch post:
- 50.3% on Terminal-Bench 4.0 at maximum reasoning effort, against 21.5% for GPT-5.6 Terra.
- On OpenAI's internal benchmark, the error rate at low reasoning effort fell from 53.6% to 16.7% compared with GPT-5.6 Terra.
- On Agents' Last Exam, a test of how well agents use software at work, OpenAI says Terra matches GPT-6 Astra at a tenth of the price per token and finishes tasks over three times faster than Astra and GPT-6.1 Sol.
Notice what each company chose to measure. Google tested hard, long coding and research jobs. OpenAI tested speed, cost and reliability on everyday agent work. That tells you more about the intended use than any single score does.
How Much Does Each Cost in the API?
Terra is cheaper today, and the gap doubles soon.
| Per 1M tokens (USD) | GPT-6.1 Terra | Gemini 4 Argon |
|---|---|---|
| Input | $1 | $2 until 30 November, then $4 |
| Output | $5 | $10 until 30 November, then $20 |
| Max output | 128K tokens | 1M tokens |
| Context window | 1.05M tokens | Not stated on the pricing page |
| Free API tier | No | No |
Terra's prices and limits come from OpenAI's model documentation, which lists $1 in and $5 out, a 1.05M token context window, 128K max output and a June 2026 knowledge cutoff. That's half the price of GPT-6.1 Sol.
Argon's are on Google's Gemini API pricing page, which shows introductory pricing of $2 in and $10 out through 30 November, rising to $4 and $20 from 1 December. It also says Argon "is not available on the Free tier," and that every token is billed, including thinking tokens. Batch and Flex are half the standard rate. On standard pricing, Argon goes from twice Terra's cost to four times.
If you want a Google model nearer Terra's price, Gemini 3.8 Flash is listed at $0.75 in and $3.75 out through the end of the year.
Which Paid Plan Gets You Each Model?
- GPT-6.1 Terra: ChatGPT Plus at $20 a month, about SGD 27, plus Pro, Enterprise and Edu. It lives in ChatGPT Work and Codex rather than the regular chat window.
- Gemini 4 Argon: Google AI Pro at SGD 28.99 a month in Singapore, and Google AI Ultra from SGD 139.99, per Google's Singapore plans page. Pro lists "Access to Gemini 4 Argon" and Ultra lists "Highest access."
Google AI Plus, the SGD 6.98 tier, doesn't include Argon. Its plan card now lists only the Flash model. ChatGPT Go doesn't include Terra either.
One practical wrinkle: Terra in ChatGPT is pitched at work tasks and Codex, so if you mostly chat, write and ask questions in the normal ChatGPT window, Plus already gives you GPT-6.1 Sol and GPT-6 Astra there. Argon, by contrast, shows up in the main Gemini app for AI Pro and Ultra, and Google lists a 1M token context window in the app on AI Pro. For long documents, that still makes AI Pro the stronger $20-class plan.
Can Free Users Try Either Model?
No. OpenAI's announcement doesn't include Free or Go, and Google's pricing page and plans page both leave Argon off the free tier.
What free users get instead is unchanged. ChatGPT Free runs GPT-6 Luna with unlimited text chats, and Gemini Free uses Auto model selection that mostly picks Gemini 3.5 Flash-Lite. Our ChatGPT vs Gemini free tier comparison covers that in detail, and ChatGPT Free remains the stronger no-cost option.
Which Should You Pick for Your Work?
Choose Gemini 4 Argon (via Google AI Pro) if you do big coding projects, long research reports, or analysis of long videos and large document sets, and you want Google's most capable consumer model. It's also the better choice if you already work in Gmail, Docs and Drive.
Choose GPT-6.1 Terra (via ChatGPT Plus) if your day is many small tasks: drafting, form-filling, spreadsheet updates, browser work or coding chores in Codex. Speed and cost per task matter more there than peak intelligence.
For developers, the maths is simple. Run high-volume agents on Terra, and save Argon for the tasks that fail on cheaper models. Budget for Argon's price rise now if you plan to lean on it.
For OpenAI's other new model against Anthropic's, see our Claude Sonnet 5.5 vs GPT-6.1 Sol comparison.
What Else Do People Ask?
Is GPT-6.1 Terra available on ChatGPT Free?
No. OpenAI says GPT-6.1 Terra is rolling out to Plus, Pro, Enterprise and Edu users in ChatGPT Work and Codex, and to developers in the API as gpt-6.1-terra. Free and Go users stay on GPT-6 Luna. The cheapest way to use Terra in ChatGPT is Plus at $20 a month, about SGD 27.
Which Google plans include Gemini 4 Argon?
Google AI Pro and Google AI Ultra. Google's Singapore plans page lists access to Gemini 4 Argon on AI Pro at SGD 28.99 a month and the highest access on AI Ultra from SGD 139.99. The free plan and Google AI Plus at SGD 6.98 don't include it.
Is Gemini 4 Argon's API price going up?
Yes. Google's Gemini API pricing page lists introductory prices of $2 per million input tokens and $10 per million output tokens through 30 November 2026, rising to $4 and $20 from 1 December 2026. Batch and Flex prices double on the same date. There's no free API tier for Argon.
Is GPT-6.1 Terra better than GPT-6.1 Sol?
Not for the hardest work. OpenAI still recommends GPT-6.1 Sol or GPT-6 Astra for the most complex coding. Terra costs half as much in the API, at $1 in and $5 out per million tokens against Sol's $2 and $10, and OpenAI says it completes tasks over three times faster than Sol and Astra.
Which model can write longer outputs?
Gemini 4 Argon, by a wide margin. Google raised Argon's output limit to 1M tokens, up from 64K. OpenAI's documentation lists a 128K token maximum output for GPT-6.1 Terra, with a 1.05M token context window. For most chats and documents, both limits are far more than you'll need.
Sources: Google's Gemini 4 Argon announcement for benchmarks, output limit and intended use. Google's Gemini app availability post for AI Pro and Ultra access. Gemini API pricing for Argon and 3.8 Flash prices. Google's Singapore plans page for plan prices and model access. OpenAI's GPT-6.1 Terra announcement for availability and positioning. OpenAI model documentation for Terra and Sol prices, context and output limits. BleepingComputer for OpenAI's reported benchmark figures. ChatGPT pricing for plan prices and free plan models. Prices and plan access change often, so check each vendor before you buy.