AI Model Value Score
Rank all 34 LLMs by quality-per-dollar. Find the cheapest model that meets your quality bar — or the most capable model within your budget.
Quality vs Cost — Scatter Plot
Premium
Mid
Budget
| # | Model | Tier | Quality | Input $/1M | Output $/1M | Avg Cost | Value Score |
|---|
How Value Score Works
Value Score = Quality Score / Average Cost per 1M tokens. Higher is better — you get more capability per dollar.
- Quality Score (50-100): Estimated from MMLU, HumanEval, MATH, and instruction-following benchmarks. Frontier models score 90+, mid-tier 80-90, budget 65-85.
- Average Cost: Mean of input and output pricing per 1M tokens. A model with $1.00 input / $5.00 output has avg cost $3.00.
- Value Score: Quality / Avg Cost. A model scoring 90 quality at $3.00 avg cost = 30.0 value score. A model scoring 75 at $0.50 avg cost = 150.0 value score.
Tip: For production workloads, set a quality threshold first (e.g. 85+), then sort by value score within that threshold to find the cheapest model that meets your bar.