THE SCOREBOARD FOR WHAT'S NEXTAI + QUANTUM, SCORED. COMPARED. EXPLAINED.
AIQUANTUMSCORE.
← AI & Quantum News

AI / NEWS BRIEF

OpenAI launches GPT-6.1 Terra at half Sol's token price

GPT-6.1 Terra brings million-token context, visual reasoning and computer use to OpenAI's mid-tier model at $1 per million input tokens and $5 per million output tokens.

By AIQuantumScore · · Event 2026-10-26

A cheaper model for agentic work

OpenAI has released GPT-6.1 Terra through its API, ChatGPT Work and Codex. The model accepts text and images, supports computer-use and visual-reasoning workflows, and has a 1.05-million-token context window. Standard API pricing is $1 per million input tokens, $0.05 per million cached input tokens and $5 per million output tokens—half the standard token price of GPT-6.1 Sol. It is available to Plus, Pro, Enterprise and Edu users in Work and Codex, but not in regular Chat.

Independent tests show a genuine mid-tier tradeoff

Artificial Analysis evaluated Terra across several reasoning settings and currently places its max-effort result below GPT-6.1 Sol on its Intelligence Index while tying Sol on that evaluator's Terminal-Bench 4.0 run. Plotly separately tested the model on 43 private data-analysis tasks: Terra completed 40 correctly for $1.31, compared with 43 for Sol at $2.61 and 38 for Luna at $0.13. Those results support the intended price-performance position, but each benchmark measures a narrow workload and should not be generalized into a universal ranking.

Provider claims remain provider claims

OUR ANALYSIS

OpenAI reports large improvements over GPT-5.6 Terra in computer use, terminal coding and factual reliability. Those launch figures were selected and run by OpenAI, so they are not presented here as independent evidence. The external tests also show that Terra can trail Sol on harder judgment and hidden-pattern tasks. Organizations choosing between them should evaluate the actual workflow, reasoning setting, latency and error cost rather than relying on one aggregate score.

Why it matters

OUR ANALYSIS

Terra materially lowers the cost of capable visual and computer-use agents without dropping to OpenAI's smallest tier. That makes it relevant for high-volume office, support and data workflows where Sol-level accuracy is useful but not always worth twice the token price. Its legitimate external benchmark observations belong in source-labelled provider data; no REVAL score should be assigned until AIQuantumScore runs the model through the private REVAL task bank.

More to follow