OpenAI released GPT-6.1 Terra on October 26, charging developers half the per-token price of its predecessor. Independent tests from Artificial Analysis, however, still rank 2 Anthropic models ahead of it.
Terra now costs $1 per million input tokens and $5 per million output tokens. That compares with $2 and $12 for GPT-5.6 Terra, according to Artificial Analysis.
Where Terra Lands on the Independent Scoreboard
Artificial Analysis, which tests artificial intelligence (AI) models, scored Terra at 48 on its Intelligence Index at maximum effort. The older GPT-5.6 Terra scored 42.
The firm put Terra’s cost per task at $0.40, against $1.40 for its predecessor. It also found Terra guesses on 38% of questions it cannot answer, down from 88%.
Anthropic’s models still hold the top of the same index. Claude Opus 5.5 leads with 58 points, while Claude Sonnet 5.5 scores 56. Google’s Gemini 4 Argon and OpenAI’s own GPT-6 Astra both score 53.
OpenAI’s launch post frames the contest around cost instead. OpenAI says Terra costs about 9 times less per task than Opus 5.5 on its own Agents’ Last Exam. That comparison pits OpenAI’s mid-range model against one of Anthropic’s pricier, higher-end models.
OpenAI also steers its hardest coding work toward pricier models.
“For the most complex coding tasks, GPT-6.1 Sol and GPT-6 Astra remain the better choice,” it said.
Investors can now set those rankings against hard financial data, because Anthropic opened its books on the same day.
A Price Cut on Anthropic’s Filing Day
Anthropic publicly filed its S-1 on October 26 and plans to list on Nasdaq under the ticker ANT. The filing shows $16.29 billion in revenue for the first half of 2026.
OpenAI has chosen a slower route to public markets. Chief executive Sam Altman has ruled out a listing this year, and the company is raising private capital instead. The Financial Times reported that OpenAI lowered the valuation target for that round to $1.25 trillion.
Terra’s pricing also changes what OpenAI earns from each task. Artificial Analysis puts Terra’s cost per task about 70% below GPT-5.6 Terra’s at maximum effort.
On those figures, OpenAI would need to sell roughly 3.5 times as many tasks to keep its revenue from that work flat. OpenAI is pitching Terra for exactly that kind of high-volume work, such as processing incoming invoices.
Anthropic’s expected mid-November debut should show how investors value a lab that tops independent scores while costing more per task.
Subscribe to our YouTube channel to watch leaders and journalists provide expert insights
