Artificial Analysis v4.2

AI Coding Subscription Value & Pareto Frontier

Efficiency Frontier (Intelligence vs Monthly Cost)
5 Frontier Anchors
X-Axis Scaling
Budget Cap Filter
Free/Contributor APIs
Benchmark Focus
Pareto Optimal Sweet Spot: Claude 3.5 / 3.7 Sonnet ($20 Cursor / Claude Pro) & DeepSeek-Coder-V2.5 API ($3) dominate the cost-per-intelligence curve.
Points on the dotted green frontier offer non-dominated coding intelligence per dollar spent.
Claude 3.7 Sonnet (Cursor Pro $20)
Pareto Optimal
Cursor / Anthropic Bundle
$20 / month
Intelligence Score
92.4 / 100
Cost / Intel Unit
$0.22
Per index point
Context Window
200K
Tokens usable
SWE-bench Ver.
70.3%
Verified fix rate
Rate Cap / Mo
500 fast
+ unmetered slow
Subscription Matrix (v4.2 Data)
14 subscriptions evaluated
Status Subscription Plan Base Model Price Score CPI
Tool Stacking & ROI Simulator
Optimal Stack Recommendation: Cursor Pro ($20) + DeepSeek PayG ($3)
Effective Monthly Spend: $23.00 / mo
Savings vs $200 Solo Pro Tier: $177.00 / mo (88.5% saved)
Enjoy this tool? Build your own with Super