Anthropic’s Claude Fable 5.2 at its max setting leads the Artificial Analysis Intelligence Index with a score of 65, according to the live leaderboard we read at 16:24 UTC on 11 November. Google’s Gemini 4 Carbon (high) scores 60 and costs $7.76 per index task, against $26.54 for Fable 5.2 (max).
By our arithmetic, Fable 5.2’s top setting costs about 3.4 times as much per task as Carbon’s for 5 more points. A post on X that put Gemini 4 Carbon first at 58.3 does not match the leaderboard, which shows Carbon (high) fourth among model settings.
Business Pill · Tokens Explained
An evergreen Business Engineer explainer on tokens, the unit AI models are priced in. It teaches the general idea only and does not discuss the index or these models.
The key insight: On the live Artificial Analysis Intelligence Index, read at 16:24 UTC on 11 November, Claude Fable 5.2 (max) leads at 65 and Gemini 4 Carbon (high) scores 60. Fable’s top setting costs $26.54 per index task against $7.76 for Carbon’s, about 3.4 times as much by our arithmetic. A viral post putting Carbon first at 58.3 does not match the leaderboard.
Three Fable 5.2 settings sit at the top: max at 65, then xhigh and high at 62 each. Gemini 4 Carbon (high) follows at 60, then Claude Opus 5.5 (max) at 59 and Gemini 4 Carbon (medium) at 58. GPT-6 Astra (max) and Claude Opus 5.5 (xhigh) score 57.
Taking the best setting of each model, the order is Fable 5.2 at 65, Gemini 4 Carbon at 60, Opus 5.5 at 59, GPT-6 Astra at 57, and GPT-6.1 Sol and Claude Sonnet 5.5 at 55. Version 5.0.1 of the index combines nine evaluations, including Terminal-Bench 4.0, HLE-Diamond and AutomationBench-AA, Artificial Analysis says.
Artificial Analysis Intelligence Index score against weighted average cost per index task (log scale) for the top 15 current model settings, read from the live leaderboard at 16:24 UTC on 11 November 2026.
What Each Point Costs
The leaderboard prices each setting by its weighted average cost per index task. Fable 5.2 (xhigh) costs $14.83 and Fable 5.2 (high) $9.08, so the high setting scores 62 at about 34% of the max setting’s cost, by our arithmetic.
Fable 5.2 (high) scores 2 points above Gemini 4 Carbon (high) for $1.32 more per task. At the low end of the top group, GPT-6.1 Sol (max) scores 55 at $1.27 a task, about 5% of Fable 5.2 (max)’s cost by our arithmetic. Our earlier pieces cover Carbon’s launch pricing and Perplexity’s Carbon test.
Time Per Task
Fable 5.2 (max) takes 45.0 minutes per index task on the leaderboard, against 26.1 for Gemini 4 Carbon (high) and 16.9 for Fable 5.2 (high). On a different measure, Epoch AI’s capabilities index, Fable 5.2 also set a record at 174.
The Structural Read
As we read the leaderboard, the top of the index is a price ladder as much as a ranking: Fable 5.2 (high) scores 62 at $9.08, close to Carbon’s $7.76, while the last 3 points to 65 cost $17.46 more per task by our arithmetic.
Artificial Analysis leaderboard summary, 11 Nov 2026
“Claude Fable 5.2 (max) and Claude Fable 5.2 (xhigh) are the highest intelligence models, followed by Claude Fable 5.2 (high) and Gemini 4 Carbon (high)”
Three Implications
FOR BUYERS Fable 5.2 (high) scores 62 at $9.08 a task, about 34% of the max setting’s cost, by our arithmetic.
FOR GOOGLE Gemini 4 Carbon (high) is the best non-Anthropic setting on the leaderboard at 60, for $7.76 a task.
FOR OPENAI GPT-6 Astra (max) scores 57 at $6.24, and GPT-6.1 Sol (max) 55 at $1.27, the leaderboard shows.
The framework says: “The cost of intelligence is collapsing while the cost of deploying intelligence is exploding.”
As we read it, the leaderboard shows both sides at once: a 55 costs $1.27 a task, while the top score costs $26.54. The framework predates these models and does not discuss the index.
Highest-scoring setting of each model on the Artificial Analysis Intelligence Index v5.0.1, with cost per index task. Source: Artificial Analysis leaderboard, read 16:24 UTC, 11 November 2026.
What Is Not Established
The leaderboard is live and its scores can change as Artificial Analysis adds models or settings. We did not find the original post that cited 58.3, so we cannot say which version of the index it used. Cost per task depends on each model’s list prices and token use in these tests, not on any customer’s bill.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
At 16:24 UTC on 11 November, the Artificial Analysis Intelligence Index put Claude Fable 5.2 (max) first at 65 and Gemini 4 Carbon (high) at 60. Fable’s top setting costs $26.54 per index task against $7.76 for Carbon’s, about 3.4 times as much by our arithmetic.
A note on sourcing. Scores, costs per task and times are from the Artificial Analysis leaderboard as we read it at 16:24 UTC on 11 November 2026; the index description is from Artificial Analysis’s homepage. The 58.3 figure is from a post on X, not the leaderboard. Ratios and differences are our arithmetic. Artificial Analysis, Anthropic and Google were not contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
