Skip to content

RFC-033: Quantitative Outcome Evaluation Model — 8. Cost Model

AIGP SpecificationRFC-033: Quantitative Outcome Evaluation Model › 8. Cost Model

← 7. Governance Actions · Section index · 9. Relationship to Existing Components →

8. Cost Model

Component Cost per 1000 invocations Notes
Deterministic (Qs, Qe) $0 Computed from existing data
LLM grading (Qc, Qk, Qr) at 10% ~$0.15 Haiku, ~500 tokens per grade
Embedding similarity (Qn) ~$0.01 Titan Embed, batch
Total (standard tier) ~$0.16 / 1000 invocations

At 10,000 invocations/day across platform: ~$1.60/day = ~$48/month.



← 7. Governance Actions · Section index · 9. Relationship to Existing Components →