Not Diamond

What is Not Diamond?

Not Diamond is a model routing service for engineering teams that predicts which LLM to use for each input and routes work to cheaper models where quality allows. It currently focuses on coding agents.

Key features of Not Diamond

• Model routing: predicts the best model per input.
• Cost control: routes to cheaper models when possible; the vendor claims about 50% cost savings and 5% accuracy gains.
• Sub-agents, caching and context: support for agent workflows.
• Secure API: works with existing harnesses and gateways to avoid vendor lock-in.
• Security: the vendor states SOC 2 and ISO 27001 compliance.
• Enterprise support: custom SLAs and 24/7 support.

How Not Diamond works

Requests pass through Not Diamond, which scores them and sends each to a suitable model from providers including Anthropic, OpenAI and open source models. Teams keep their existing tooling and see savings from routing.

Who uses Not Diamond?

It is used by engineering organizations with heavy coding agent usage that want to lower model spend. The vendor models savings on a sample team of 1,000 engineers.

Not Diamond pricing

Not Diamond does not list plans on its homepage; demos and custom plans are offered through its Book a demo page.

Not Diamond integrations

The vendor names Anthropic, OpenAI and open source models as supported providers and works with existing harnesses and gateways.

Not Diamond alternatives

Langfuse is an observability and evaluation platform for LLM apps rather than a router. PromptLayer manages and monitors prompts and requests. Vellum is a platform to build and evaluate LLM workflows.