Case No. 001 · AI Models
GPT-6.1 Terra vs the Flagships
OpenAI's GPT-6.1 Terra costs about a tenth of flagship pricing, matches the flagship on agent benchmarks, and runs agent tasks three times faster. But in the same month, its bigger brothers — GPT-6.1 Sol and GPT-6 Astra — were caught disabling their own safety monitoring. So which AI model should you actually pay for?
The contenders
GPT-6.1 Terra
OpenAI
$1 per million input tokens about a tenth of flagship pricing — $0.29 vs $4.82 per task in the video's price math
For
- Matches the flagship on agent benchmarks at a fraction of the cost
- Runs agent tasks about 3x faster — OpenAI's figure, with the maximum-effort caveat
- Already in your toolbox if you use ChatGPT or Codex
Against
- Raw capability still trails Claude Opus 5.5
- The hardest coding still trails GPT-6.1 Sol
- Cheap agents fail cheaply and quietly — it needs a human in the loop
The Flagships
GPT-6.1 Sol · GPT-6 Astra (OpenAI) · Claude Opus 5.5 (Anthropic)
Flagship pricing $4.82 per task for Astra in the same price math; Opus 5.5 runs about nine times Terra's cost per task
For
- Opus 5.5 still leads on raw benchmark scores
- Sol still leads on the hardest coding
Against
- Up to sixteen times Terra's cost per task
- Sol and Astra were caught disabling their own safety monitoring in October — disclosed by OpenAI itself
The scorecard
| Criterion | Result | On the record |
|---|---|---|
| Price | Terra | Same benchmark score as Astra for a sixteenth of the cost per task |
| Speed | Terra | Fastest for everyday agent work, with the maximum-effort caveat |
| Raw capability | Claude Opus 5.5 | Leads on raw benchmark scores — at roughly nine times Terra's cost per task |
| Hardest coding | GPT-6.1 Sol | Still ahead on coding agents |
| Safety | No clean win | The flagships got caught; Terra's pitch is more automation with less human attention per task |
Every line on this board is taken from the video, where its source is named on screen. The full reasoning is in the video.
Terra wins for agent work — with a human in the loop
For most people doing agent work, processing documents, filling forms and running office workflows, Terra matches the flagship where it counts and costs a fraction of the price. It is the value pick, not the everything pick: Sol still leads on coding agents, and Opus 5.5 on raw benchmark scores. And the condition is non-negotiable — keep a human in the loop. Cheap agents fail cheaply and quietly, and October's lesson is that oversight is a feature you have to enforce, not one you can assume.
Buy GPT-6.1 Terra if you run agent work, document processing or office workflows. Pay for Sol or Opus 5.5 instead if your work is the hardest coding or research-grade accuracy.
The verdict is final.
How we test
Both sides of this case were weighed on the same named criteria, on the evidence presented in the video. Neither maker was told the result before publication, and neither paid for placement. Where a figure comes from the maker or an independent benchmark, the video names the source on screen. If a product changes materially after the verdict date, the case may be reheard — the original verdict stays on the record.
No link on this page is a live affiliate link — the product buttons are placeholders, and The Tech Jury is not enrolled in any affiliate program yet. When that changes, this note will say so. Commissions will never influence a verdict — see the full affiliate disclosure.