Driver Analysis
Analyze the key factors that drive AI model rankings. Drivers represent the specific signals that most influence where a model lands in the leaderboard -- both positive boosters and negative detractors.
Unique Drivers
6
Models with Drivers
300
Top Positive Driver
Pricing
280 models
Top Negative Driver
Capabilities
4 models
Top Drivers by Frequency
Driver Frequency
How often each driver appears across all ranked models, broken down by impact type.
| Driver | Total | Net Impact |
|---|---|---|
| Capabilities | 297 | +240 |
| Pricing | 281 | +280 |
| Recency | 234 | +232 |
| Benchmarks | 225 | +196 |
| Context Window | 143 | +143 |
| Output Capacity | 20 | +20 |
Positive Drivers - What Helps
The top 6 most common positive-impact drivers that boost model rankings.
Pricing
280 models$25.00/M output tokens
Capabilities
244 modelsSupports reasoning, vision, tools, JSON mode, web search, streaming
Recency
232 modelsReleased 2 months ago
Benchmarks
197 modelsContext Window
143 models1M token context window
Output Capacity
20 modelsUp to 128K output tokens per request
Negative Drivers - What Hurts
The top 2 most common negative-impact drivers that push model rankings down.
Capabilities
4 modelsSupports reasoning, vision, tools, JSON mode, web search, streaming
Benchmarks
1 modelsDriver Breakdown by Model
Top 20 models by score with their individual driver breakdown.
Driver Signal Distribution
Drivers grouped by their underlying signal category, showing the distribution of positive, negative, and neutral impacts.
| Signal | Count |
|---|---|
| capability | 297 |
| pricing_tier | 281 |
| recency | 234 |
| benchmark | 225 |
| context_window | 143 |
| output_capacity | 20 |
Methodology
How driver analysis works.
What Drivers Represent
Drivers are the specific factors that most influence a model's position in the leaderboard. Each driver captures a distinct aspect of model quality, pricing, capabilities, or market performance that contributes to the composite ranking score.
How They Are Computed
Drivers are derived from the scoring algorithm that evaluates models across multiple dimensions. The algorithm identifies which signals have the greatest impact on each model's final ranking, then surfaces the top contributors as drivers with their corresponding impact direction and metric values.
Impact Types
Positive drivers help a model rank higher -- the model excels in this area. Negative drivers push a model down -- this is an area of weakness. Neutral drivers are present but do not significantly affect ranking in either direction.
Explore More
Dive deeper with signal analysis, benchmark comparisons, or browse all explorer tools.
Ranking drivers are the individual factors that push a model's composite score up or down. Positive drivers (like strong benchmark performance or competitive pricing) boost a model's rank, while negative drivers (like limited capabilities or high cost) pull it down.
Each driver represents the difference between a model's signal score and the average across all models, weighted by importance. Signals include capability breadth, pricing tier, context window size, recency, output capacity, and versatility.
Benchmark performance accounts for 90% of the score, from MMLU, GPQA, HumanEval, SWE-bench, and 15+ standardized evaluations. Capabilities and context window serve as tiebreakers (10%), while output capacity and versatility each add 10%.