Pricing for AI teams shipping production AI.
The deterministic alternative to LLM-as-judge pricing. Context-aware evaluation, 49 dimensions, and production A/B testing — no judge-model surcharge, no token-cost surprises. Pay for what you use.
$0
forever free
Get started- 1 project
- 1 experiment, 3 variants
- Community support
$49
/month
Get started- 5 projects
- 10 feature flags per project
- Email support (24hr)
$249
/month
Get started- Max-Accuracy mode
- Document review
- 25 projects
- Unlimited feature flags
- Email + Slack support
Custom
tailored to your needs
Contact sales- Self-hosted / VPC deployment
- Unlimited projects
- Custom SEU caps
- SSO/SAML authentication
- RBAC & audit logs
- Dedicated support
Standard Evaluation Unit
One SEU covers one evaluation across all 49 dimensions, with up to 10,000 claim-tokens of grounding — the number of claims in the answer times the reference tokens they are checked against (for example, 5 claims against 2,000 tokens). A typical RAG or chat answer is 1 SEU. Longer answers and larger references use more, in proportion to the checking they need.
Every mode costs the same. You pay for how much is checked, never for which feature checked it.
| One evaluation, all 49 dimensions, any mode, up to 10,000 claim-tokens | 1 SEU |
| Each further 10,000 claim-tokens | +1 SEU |
Bring your own key (BYOK)
You connect your own LLM API keys and pay the providers directly. variA/Bly charges a flat platform fee per SEU — never a markup on model tokens.
- Platform fee: from $0.009/SEU
- Use your existing API contracts
- Full control over LLM spend
Save with annual plans
Get 2 months free (~17% discount) with annual prepayment.
Ready to ship AI you can defend?
Deterministic evaluation gives you the audit trail behind every response.