Understand this story
On October 20, 2026, Artificial Analysis put forward the judgment that the cost of reaching each score threshold on the Intelligence Index is falling rapidly. Based on v4.3.2 measured results and current pricing statistics, it stated that the per-task cost of reaching 30 points is 86 times lower than when that score was first reached: GPT-6 Luna (high) reaches 30 points at $0.03 per task, whereas Claude Sonnet 4.6 cost $2.49 per task when it first reached 30 points in February 2026. The same statistics also stated that the cost of reaching 20 points is 123 times lower than about 13 months ago, and the cost of reaching 40 points is 33 times lower than about 5 months ago; Claude Haiku 5.5 (xhigh) reaches 41 points at $0.12 per task, making it the cheapest model to reach 40 points. The highest measured score rose from 21 to 58 within a year, at a cost of about 11 times, with Claude Opus 5.5 at $5.98 per task. At present, this judgment is published by Artificial Analysis as a single source, and no other independent source has cross-verified it.
AI-generated from public reports
Developments
1 confirmed developments
Artificial Analysis says the cost of reaching the intelligence index threshold is falling rapidly
Source:Artificial Analysis
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 21
Artificial Analysis
Artificial Analysis: Reaching an Intelligence Index score of 30 is now 86 times cheaper than when it was first achieved
Based on v4.3.2 measured results and current pricing, Artificial Analysis found that the per-task cost of reaching an Intelligence Index score of 30 is 86 times lower than when that score was first achieved. GPT-6 Luna (high) reaches 30 at $0.03/task, whereas Claude Sonnet 4.6 cost $2.49/task when it first reached 30 in February 2026. Reaching 20 is about 123 times cheaper than roughly 13 months ago, and reaching 40 is about 33 times cheaper than roughly 5 months ago. Claude Haiku 5.5 (xhigh) reaches 41 at $0.12/task, making it the cheapest model to reach 40. The highest measured score rose from 21 to 58 within a year, at roughly 11 times the cost, with Claude Opus 5.5 at $5.98/task.