Pricing
Free tier math
Included access uses simple daily checks.
One check means one prompt sent to one model. That keeps the quota easy to understand and makes mixed runs predictable.
1 prompt x 10 models = 10 checks
2 prompts x 10 models = 20 checks
5 prompts x 4 models = 20 checks
If you already connected your own OpenAI, Anthropic, or Google key, ModelGrade automatically falls back to your own key after the free included quota is used up.
Current status
Buying is under development
For now, start on the free tier or connect your own provider keys. The paid quota upgrade flow is the next billing step, but it is not purchasable from the app yet.
Free
Free
Best for POC and first evaluation loops.
100 included checks per day
Each prompt x model pair counts as one check
Bring your own provider keys anytime
If your own key is connected, ModelGrade automatically uses it after the included quota is exhausted
Pro
$99/mo
Higher daily quota for regular team usage.
Higher included daily quota
Better team capacity and repeated benchmarking
Same bring-your-own-key fallback behavior
Ideal once the free quota becomes a daily bottleneck
Enterprise
Custom
For larger workloads and deployment support.
Custom quota and support
Larger scale benchmarking and routing workflows
Deployment guidance
Best for shared internal evaluation programs