Blended Model Cost Calculator
Estimate cost when routing traffic between two models.
With the default inputs, $0.00161 blended per request — $161.00/month routing 80% to GPT-4o mini. Enter your own numbers below to recompute instantly; the full step-by-step math is shown under the worked example.
Results update automatically as you type.
- GPT-4o mini / request
- $0.000450
- GPT-5 / request
- $0.00625
- Blended / request
- $0.00161
- Blended / month
- $161.00
Many apps route easy requests to a cheap model and hard ones to a premium model. Set the split to see your blended per-request and monthly cost.
How this is calculated
We cost each request on both models, then blend by your routing split: blended cost = %A × cheap-model cost + %B × premium-model cost, scaled by requests per month. It shows what routing most traffic to a cheap model saves versus using the premium model for everything.
Is this a good result? What to do next
Routing pays off when the cheap model handles the majority of requests well, because the premium model only ever sees the hard minority. The savings can be large, but the split must be validated against quality, not chosen to hit a cost target.
Typical planning ranges
- 80/20 split
- large saving vs 100% premium
- Savings track
- the cheap-model share
- Safe split set by
- quality, not cost
Ranges are typical planning figures to sanity-check your result, not authoritative benchmarks. Your numbers will vary with use case, volume, and vendor.
How to improve this number
- Raise the cheap-model share as far as quality allows.
- Add a classifier or confidence threshold to route requests.
- Escalate to the premium model only on low confidence.
Common mistakes
- Setting the split by guess instead of measured quality.
- Ignoring the overhead of the routing/classifier step.
When to use a different approach
To pick the two models, use a two-model comparison like GPT vs Claude or the model cost comparison. For a single model, use the LLM API cost calculator.
Worked example (defaults)
With the default inputs above, here is the result:
- GPT-4o mini / request
- $0.000450
- GPT-5 / request
- $0.00625
- Blended / request
- $0.00161
- Blended / month
- $161.00
- blended = 80% x $0.000450 + 20% x $0.00625
Sources & references
Frequently asked questions
What is model routing?+
Sending each request to the cheapest model that can handle it, escalating to a stronger model only when needed.
How do I pick the split?+
Start from your data on how often the cheap model is sufficient, then tune based on quality metrics.