OpenAI's Pitch for GPT-6.1 Terra: Lower Price, Not Higher Scores

On October 26, OpenAI rolled out GPT-6.1 Terra, pricing it at half the per-token rate of the earlier model. Even so, Artificial Analysis' independent evaluations continue to place 2 Anthropic models above it.

$1 per million input tokens and $5 per million output tokens are Terra's current rates. For GPT-5.6 Terra, the comparable figures were $2 and $12, according to Artificial Analysis.

Terra's Standing on the Independent Scoreboard

The AI-testing firm Artificial Analysis gave Terra 48 points on its Intelligence Index at maximum effort. The earlier GPT-5.6 Terra registered 42.

Terra's per-task cost is put at $0.40, versus $1.40 for the older release. On questions Terra cannot answer, it resorts to a guess 38% of the time, down from 88%, the firm found.

The top of that index remains in Anthropic's hands: Claude Opus 5.5 heads the list with 58 points, and Claude Sonnet 5.5 follows with 56. Google's Gemini 4 Argon and OpenAI's own GPT-6 Astra both sit at 53.

OpenAI's launch post puts the emphasis on cost instead. Terra, OpenAI says, works out to about 9 times less per task than Opus 5.5 on its own Agents' Last Exam. That comparison pairs OpenAI's mid-range model with one of Anthropic's pricier, higher-end systems.

OpenAI also directs its most complex coding work toward the more expensive models.

“For the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice,” OpenAI said.

Because Anthropic opened its books that same day, investors can now judge those rankings against concrete financial data.

On Anthropic's Filing Day, Terra Gets Cheaper

Anthropic made its S-1 public on October 26 and intends to list on Nasdaq under the ticker ANT. The S-1 reports $16.29 billion in revenue for the first half of 2026.

OpenAI is moving more slowly toward public markets. CEO Sam Altman has said a listing is off the table this year, and private capital is being raised instead. OpenAI cut the valuation target for that round to $1.25 trillion, the Financial Times reported.

Terra's pricing also alters the revenue OpenAI earns per task. Artificial Analysis has Terra's per-task cost about 70% below GPT-5.6 Terra's at maximum effort.

OpenAI would have to sell roughly 3.5 times as many tasks to hold that revenue stream flat, based on those figures. Terra is being pitched for precisely that type of high-volume job, like handling incoming invoices.

Anthropic's debut, expected in mid-November, should reveal how investors value a lab that leads independent scores while charging more per task.