HomeDemandPricing Tiers
Price per Intelligence Point and Cost per Task
By list price
Intelligence Index vs Blended Price (Snapshot Sep 4 2026)#
Fable 5.1 pushed the top of the frontier to 65.7 points at the same $20 list price; GPT-6 Astra costs the same $20 for 61 points.Draft, pending review
How to read this chart
One dot per model (intelligence score 20 or above, 246 models). Color is launch period; dashed lines are the cumulative price-performance frontier for each period (list-price basis).
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-01
Cheapest Way to Buy the Same Capability, Over Time (log)#
The cheapest price for the same capability keeps stepping down; the 60-plus line fell 90% in three months.Draft, pending review
How to read this chart
The lowest blended price on record to reach intelligence scores of 30, 40, 50 and 60, by launch date at current prices. The 60-plus line: $20 in June, $8 in July, $3 in August, $2 in September.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-02
By cost per task
Intelligence Index vs Cost per Task (Cost to Run the Intelligence Index, Snapshot Sep 4 2026)#
Measured by cost per task, China's small models took over the cheap end of the frontier: capability that cost $779 in late 2025 cost $25 by 1H26.Draft, pending review
How to read this chart
The x-axis is the actual cost of running the full intelligence benchmark, including reasoning tokens, on a log scale; top-left is strong and cheap. Dashed lines are cumulative frontiers by period.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-03
How Chatty Each Model Is: Tokens Burned Running the Same Benchmark (Implied, Millions)#
On the same benchmark, Sonnet 5 at max effort burns a billion tokens, 31 times the most frugal model; GPT-6 Astra is the most frugal in the $20 band.Draft, pending review
How to read this chart
Implied consumption = task cost ÷ list price, for the 17 models scoring 40 or above. Red is China, blue is the US.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-04
Task Cost per Intelligence Point (USD per Point, log)#
On a per-task basis the ranking flips: Fable 5.1 at max effort is the priciest at $130 per point, paying for the last three points nobody else has.Draft, pending review
How to read this chart
Task cost divided by intelligence score; compare with the list-price-per-point ranking, which flips substantially.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-05
Real-world benchmark: CursorBench
CursorBench 3.2: Agentic Coding Score × Cost per Task (Effort Sweep, Snapshot Sep 2 2026)#
On the real-world coding benchmark Fable 5.1 took first place while costing about half as much per task as its predecessor.Draft, pending review
How to read this chart
Cursor's official benchmark, built from real multi-file, loosely specified tasks. X is actual cost per task (log), y is score; dots of one color on one line are the same model swept across effort settings from low to high.
Source: Cursor public benchmarks; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-06
Same Job, Who Talks Most: Tokens per Task (Top Effort Setting per Model, CursorBench 3.2, Snapshot Sep 2 2026)#
The efficiency gain came from talking less, not from a price cut: Fable 5.1 kept its list price, used 30% fewer tokens, and cost 40% less.Draft, pending review
How to read this chart
Each model at its top effort setting. Red is China, blue is the US, brown is Cursor's own model. Cost per task = tokens × list price, so a cheap but talkative model is not necessarily cheap.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-07
List price per point of intelligence
What Intelligence Costs on Today's Frontier Models (USD per Million Tokens per Intelligence Point)#
Per point of intelligence, Fable 5.1's list price is 12 times Gemini 3.8 Flash's: the flagship premium buys the few points nobody else can deliver.Draft, pending review
How to read this chart
List price per million tokens divided by intelligence score, not a per-task measure; lower is better value.
Source: Published model pricing and the Artificial Analysis index; FinSight compilation · Updated 2026-09-04
Standalone page ↗d3-intel-price-mon-08