File size: 8,270 Bytes
82697cc |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 |
#!/usr/bin/env python3
"""
OpenRouter API Pricing Analysis
Analyzes context windows, token pricing, and output/input ratios
"""
import json
import numpy as np
from pathlib import Path
from typing import Dict, List, Any
import statistics
def load_data(file_path: str) -> List[Dict[str, Any]]:
"""Load the models data from JSON file"""
with open(file_path, 'r') as f:
return json.load(f)
def calculate_distribution_stats(values: List[float], name: str) -> Dict[str, Any]:
"""Calculate comprehensive distribution statistics"""
if not values:
return {"error": "No data"}
values = sorted(values)
n = len(values)
stats = {
"name": name,
"count": n,
"min": min(values),
"max": max(values),
"mean": statistics.mean(values),
"median": statistics.median(values),
"std_dev": statistics.stdev(values) if n > 1 else 0,
}
# Quartiles
stats["q1"] = np.percentile(values, 25)
stats["q3"] = np.percentile(values, 75)
stats["iqr"] = stats["q3"] - stats["q1"]
# Additional percentiles
stats["p10"] = np.percentile(values, 10)
stats["p90"] = np.percentile(values, 90)
stats["p95"] = np.percentile(values, 95)
stats["p99"] = np.percentile(values, 99)
return stats
def calculate_stratification(values: List[float], name: str) -> Dict[str, Any]:
"""Calculate stratification by creating bins/ranges"""
if not values:
return {"error": "No data"}
values_array = np.array(values)
# Determine appropriate bins based on data range
min_val = values_array.min()
max_val = values_array.max()
# Create stratification bins
if max_val - min_val < 1:
# For very small ranges (like some pricing)
bins = np.linspace(min_val, max_val, 11)
else:
# Use log scale for wide ranges
if min_val > 0:
bins = np.logspace(np.log10(max(min_val, 0.0001)), np.log10(max_val), 11)
else:
bins = np.linspace(min_val, max_val, 11)
hist, bin_edges = np.histogram(values_array, bins=bins)
stratification = {
"name": name,
"bins": []
}
for i in range(len(hist)):
stratification["bins"].append({
"range": f"{bin_edges[i]:.6f} - {bin_edges[i+1]:.6f}",
"count": int(hist[i]),
"percentage": float(hist[i] / len(values) * 100)
})
return stratification
def analyze_context_windows(models: List[Dict[str, Any]]) -> tuple:
"""Analyze context window distribution"""
context_lengths = [
m['context_length']
for m in models
if m.get('context_length') and m['context_length'] > 0
]
distribution = calculate_distribution_stats(context_lengths, "Context Windows")
stratification = calculate_stratification(context_lengths, "Context Windows")
return distribution, stratification
def analyze_input_pricing(models: List[Dict[str, Any]]) -> tuple:
"""Analyze input token pricing distribution"""
input_prices = [
m['prompt_price_per_1m_tokens']
for m in models
if m.get('prompt_price_per_1m_tokens') is not None and m['prompt_price_per_1m_tokens'] > 0
]
distribution = calculate_distribution_stats(input_prices, "Input Token Pricing (per 1M tokens)")
stratification = calculate_stratification(input_prices, "Input Token Pricing")
return distribution, stratification
def analyze_output_pricing(models: List[Dict[str, Any]]) -> tuple:
"""Analyze output token pricing distribution"""
output_prices = [
m['completion_price_per_1m_tokens']
for m in models
if m.get('completion_price_per_1m_tokens') is not None and m['completion_price_per_1m_tokens'] > 0
]
distribution = calculate_distribution_stats(output_prices, "Output Token Pricing (per 1M tokens)")
stratification = calculate_stratification(output_prices, "Output Token Pricing")
return distribution, stratification
def analyze_output_input_ratio(models: List[Dict[str, Any]]) -> tuple:
"""Analyze output/input token price ratio"""
ratios = []
for m in models:
input_price = m.get('prompt_price_per_1m_tokens', 0)
output_price = m.get('completion_price_per_1m_tokens', 0)
# Calculate ratio only for non-free models
if input_price > 0 and output_price > 0:
ratio = output_price / input_price
ratios.append(ratio)
distribution = calculate_distribution_stats(ratios, "Output/Input Price Ratio")
stratification = calculate_stratification(ratios, "Output/Input Price Ratio")
return distribution, stratification
def format_distribution_report(dist: Dict[str, Any]) -> str:
"""Format distribution statistics as readable text"""
if "error" in dist:
return f"Error: {dist['error']}"
report = f"""
{dist['name']}
{'=' * len(dist['name'])}
Sample Size: {dist['count']:,}
Central Tendency:
Mean: {dist['mean']:.6f}
Median: {dist['median']:.6f}
Spread:
Min: {dist['min']:.6f}
Max: {dist['max']:.6f}
Std Dev: {dist['std_dev']:.6f}
IQR: {dist['iqr']:.6f}
Quartiles:
Q1 (25th percentile): {dist['q1']:.6f}
Q2 (50th percentile): {dist['median']:.6f}
Q3 (75th percentile): {dist['q3']:.6f}
Additional Percentiles:
10th: {dist['p10']:.6f}
90th: {dist['p90']:.6f}
95th: {dist['p95']:.6f}
99th: {dist['p99']:.6f}
"""
return report
def format_stratification_report(strat: Dict[str, Any]) -> str:
"""Format stratification as readable text"""
if "error" in strat:
return f"Error: {strat['error']}"
report = f"""
{strat['name']} - Stratification
{'=' * (len(strat['name']) + 18)}
Range Distribution:
"""
for bin_data in strat['bins']:
report += f" {bin_data['range']:>40s}: {bin_data['count']:>6,} models ({bin_data['percentage']:>5.1f}%)\n"
return report
def main():
# Load data
data_file = Path(__file__).parent.parent / 'raw' / 'models.json'
models = load_data(data_file)
print(f"Loaded {len(models)} models\n")
# Create output directory
output_dir = Path(__file__).parent
output_dir.mkdir(exist_ok=True)
# Analyze each metric
analyses = {
'context_windows': analyze_context_windows(models),
'input_pricing': analyze_input_pricing(models),
'output_pricing': analyze_output_pricing(models),
'output_input_ratio': analyze_output_input_ratio(models)
}
# Generate reports for each analysis
for metric_name, (distribution, stratification) in analyses.items():
# Create individual report files
dist_report = format_distribution_report(distribution)
strat_report = format_stratification_report(stratification)
# Write distribution report
dist_file = output_dir / f"{metric_name}_distribution.txt"
with open(dist_file, 'w') as f:
f.write(dist_report)
print(f"Created: {dist_file}")
# Write stratification report
strat_file = output_dir / f"{metric_name}_stratification.txt"
with open(strat_file, 'w') as f:
f.write(strat_report)
print(f"Created: {strat_file}")
# Also save as JSON for programmatic access
json_file = output_dir / f"{metric_name}_stats.json"
with open(json_file, 'w') as f:
json.dump({
'distribution': distribution,
'stratification': stratification
}, f, indent=2)
print(f"Created: {json_file}")
# Create comprehensive summary report
summary_file = output_dir / "summary_report.txt"
with open(summary_file, 'w') as f:
f.write("OpenRouter API Pricing Analysis - Summary Report\n")
f.write("=" * 60 + "\n\n")
f.write(f"Dataset: {len(models)} models analyzed\n\n")
for metric_name, (distribution, stratification) in analyses.items():
f.write("\n" + "=" * 60 + "\n")
f.write(format_distribution_report(distribution))
f.write("\n")
f.write(format_stratification_report(stratification))
f.write("\n")
print(f"\nCreated: {summary_file}")
print("\nAnalysis complete!")
if __name__ == "__main__":
main()
|