Intelligent
Model Routing
Automatically route every request to the optimal model. Save up to 97% on LLM costs without sacrificing quality.
Set model: "auto" — We Handle the Rest
Every request is analyzed in real-time to find the best model for cost, quality, and performance
# Just set model to "auto" - intelligence takes over
response = client.chat.completions.create(
model="auto", # Costbase picks the optimal model
messages=[{role: "user", content: prompt}],
)
# Simple query? → GPT-4o-mini ($0.15/1M tokens)
# Complex reasoning? → Claude Sonnet ($3/1M tokens)
# Coding task? → GPT-4o ($2.50/1M tokens)Simple Query
"What's the capital of France?"
94% saved
vs GPT-4o ($2.50/1M)
Complex Analysis
"Analyze Q3 financial trends..."
Optimal choice
Quality + cost balanced
Three Routing Strategies
Choose the strategy that fits your use case, or let us pick automatically
Cost-Based
Analyzes query complexity and routes to the cheapest model that meets your quality threshold. Perfect for high-volume, cost-sensitive workloads.
Typical savings
40-60%
Semantic
Uses embeddings to understand query intent and routes based on task type. Coding to GPT-4o, creative to Claude, translation to Gemini.
Typical savings
30-50%
Performance
Optimizes for latency and throughput based on real-time provider benchmarks. Ideal for user-facing applications.
Typical savings
20-30%
Route Across 8 Providers, 96+ Models
Use your own API keys. We add intelligence, you keep control.
Why Intelligent Routing?
Up to 97% Cost Reduction
Stop overpaying for simple queries with expensive models
Quality Preserved
Never sacrifice quality—we pick models that meet your standards
Zero Code Changes
Just change model to "auto" and routing intelligence kicks in
Full Transparency
See exactly which model handled each request and why
Start Saving with Intelligent Routing
One line of code. Up to 97% savings. Zero quality compromise.