Not Diamond recommends which model to call (you typically keep your own gateway and provider bills). o10 is an OpenAI-compatible inference gateway that serves tokens in the path, enforces eval floors and budget envelopes, and issues shadow/enforce savings receipts. o10 Tune is free with compute bundled.
Cost dashboards tell you what you spent. o10 sits in the request path and changes what you spend — shadow first, then enforce.
Short, self-contained answers with cited stats — read the sections below for full context.
Not Diamond is a recommendation layer for model selection. o10 is an in-path OpenAI-compatible gateway that serves tokens, enforces quality floors and budgets, and returns savings receipts.
Tune free · compute bundled
Yes if you need in-path enforcement and free prompt optimization with receipts — not only a recommendation about which model to call next.
Teams searching for a Not Diamond alternative usually need either smarter recommendations or an in-path control plane.
Not Diamond’s public positioning (notdiamond.ai) emphasizes intelligent model routing recommendations that integrate with your existing gateway/harness — it states it is not a gateway. Public pricing (notdiamond.ai/pricing, checked 2026-07) lists pay-as-you-go at about $0.05 per million tokens routed, plus enterprise options.
o10 is the gateway: base URL https://app.o10.io/v1, free route key, shadow → enforce, ledger, KYI, and Tune (free prompt optimization with held-out quality on your examples).
A fixed ~$20-per-successful-optimization fee is not currently listed on Not Diamond’s public pricing page — needs operator confirmation before citing as a hard number.
Factual rows sourced from each product’s public positioning.
Sources: notdiamond.ai, notdiamond.ai/pricing, www.o10.io, app.o10.io.
| Dimension | Not Diamond | o10 |
|---|---|---|
| Layer | Recommendation / routing intelligence (not a gateway) | In-path OpenAI-compatible gateway |
| Who serves tokens | Your providers / harness | o10 fulfills routed traffic across venues |
| Model choice | Router recommends model (+ reasoning effort) | Eval-gated pick against your quality floor |
| Spend control | Cost objectives in recommendations | Budget envelopes + shadow/enforce in path |
| Prompt optimization | Joint prompt optimization (feature; fees vary — verify) | Tune: free; failures never charge |
| Proof | Savings dashboard (their framing) | Shadow receipts + immutable ledger |
| Public fee signal | ~$0.05 / M tokens routed (public page, 2026-07) | Free BYOK start; paid plans add routed credits |
Free prompt optimization. Submit a prompt and 5–20 real examples; o10 searches cheaper models and prompt variants and returns a tuned prompt + recommended model with a before/after receipt. Quality is measured on a held-out split of your examples. Successful tunes are free; failures never charge.
Recommendation-layer tools often charge for routing and/or bill optimization compute separately (Not Diamond's public pricing lists ~$0.05 per million tokens routed as of 2026-07; a fixed ~$20-per-success optimization fee is not listed on their public pricing page — needs operator confirmation before citing). o10 Tune bundles the optimization loop on low-cost routed supply, so successful tunes are free and failures never charge.
Start free at app.o10.io/signup — free route key, no card. Point OpenAI-compatible clients at https://app.o10.io/v1.
Create an org at https://app.o10.io/signup — no card.
Mirror traffic; prove compliant savings on your workloads.
Open app.o10.io/tune with 5–20 examples; keep the receipt.
Comparison uses Not Diamond public site/pricing (2026-07) and o10 product facts. Competitor fees change — verify before citing.
Yes for teams that need an in-path gateway with quality floors, budgets, receipts, and free Tune — not only model recommendations into an existing gateway.
Not Diamond’s own FAQ states it is not a gateway and integrates with your existing gateway infrastructure. o10 is the OpenAI-compatible base URL in the path.
Public pricing at notdiamond.ai/pricing (checked 2026-07) lists ~$0.05 per million tokens routed for pay-as-you-go. Enterprise is sales-led. A ~$20-per-success optimization fee is not listed on that public page — confirm with Not Diamond before citing.
Free prompt optimization. Submit a prompt and 5–20 real examples; o10 searches cheaper models and prompt variants and returns a tuned prompt + recommended model with a before/after receipt. Quality is measured on a held-out split of your examples. Successful tunes are free; failures never charge.
Paste a week of traffic. Get the number that books the audit.
See what you're overpaying →