Total Score
0.0
Avg IQ Index 0.0
Avg Value Index 0.0

Total Score
0.0
Avg IQ Index 0.0
Avg Value Index 0.0

Leaderboard

Benchmarks Included

Intelligence vs. Value Matrix

Visualizing the trade-off between raw capability (IQ Index) and cost-efficiency (Value Index). The ideal model sits in the top-right quadrant.

USA Models
China Models
• Size = Unified Score
🏆 High IQ, High Value

The sweet spot: Smart AND affordable

💎 High IQ, Low Value

Premium frontier models

Scoring Methodology

Unified Power Score (Max 200)

The ultimate metric using Intelligence-Gated Value — where intelligence is the foundation and value acts as a multiplier.

Unified = IQ × (1 + Value / 100)

A model with zero intelligence scores zero — no matter how cheap. Value without intelligence is meaningless.

Intelligence Index ($I$)

Unweighted average (Max 100) of 11 frontier benchmarks:

  • AIME 2025: Math Olympiad (No-tool)
  • HMMT 2025: Harvard-MIT Math Tournament
  • GPQA Diamond: PhD-level science reasoning
  • ARC-AGI: Abstraction & Reasoning Corpus
  • BrowseComp: Web agent capabilities
  • ARC-AGI v2: Advanced general intelligence
  • HLE: Humanity's Last Exam
  • MMLU-Pro: Robust multi-task understanding
  • LiveCodeBench: Live coding challenges
  • SWE-Bench Verified: Real-world software engineering
  • CodeForces: Competitive programming

Value Index ($V$)

Log-normalized efficiency score (Max 100) based on Blended Cost per 1M tokens.

V = 100 × (1 - log(Cost / 0.25) / log(60.00 / 0.25))

Floor ($0.25): DeepSeek-V3.2 (Max Score) Ceiling ($60.00): Legacy Models (Min Score)

National Score Calculation

The National Score for Team USA and Team China is the sum of the Unified Power Scores for all models belonging to that nation that appear in the Global Top 10.

* Only models ranked in the top 10 globally contribute to the national total. This rewards both peak performance and depth of dominance.

Not affiliated with any model provider. No guarantees are made about correctness, completeness, or fitness for any purpose.