Company dedicated to model routing for enterprises. It can save money and often improves quality. The creator discloses being a small investor.
Not Diamond
by Not Diamond
Intelligent model router API that picks the model and reasoning effort per step for coding agents to cut inference cost.
Updated EU status checked Last shown
Not Diamond at a glance
- What is Not Diamond?
- Not Diamond is an api from Not Diamond. Intelligent model router API that picks the model and reasoning effort per step for coding agents to cut inference cost.
- Who makes Not Diamond?
- Not Diamond is made by Not Diamond.
- Does Not Diamond work in the EU?
- Unverified. No explicit evidence found. The terms say the owner is US-based, that access from outside the US is at the user's own initiative, and that there is no country exclusion list. No EU data residency is stated, though a privacy-preserving routing mode is offered. Last verified 14 Nov 2026.
- How much does Not Diamond cost?
- Paid. Pay-as-you-go at $0.05 per million tokens routed; Enterprise with volume-based discounts (custom pricing). Licence: Proprietary.
- Where was Not Diamond shown?
- In 2 creator videos by Matthew Berman. Most recently by Matthew Berman on 7 Jul 2026, at 14:53 in “Cut your AI cost IN HALF (EASY)”.
Sources: the vendor's own pages (linked above) and creator videos at the linked timestamps. EU status is information, not legal advice; see how we check.
Not Diamond is a model-routing service. For each request in a coding-agent session, it predicts which model and reasoning effort to use, with the aim of keeping output quality close to frontier models at lower cost. It is not a gateway: recommendations are executed in the customer's existing gateway and agent harness.
Routing decisions draw on payload semantics, KV cache state, implicit feedback signals, session outcomes, sub-agent structure, compaction events and prior turns, evaluated against a chosen model pool and cost objective. The vendor says it uses reinforcement learning to improve from feedback and reports at least 20-40% cost savings; these are vendor claims. An optional privacy-preserving mode routes on derived metadata and locally computed features, so the vendor does not see payloads.
Access is via API. Pay-as-you-go pricing suits developers and small teams, and an Enterprise tier adds volume discounts, SAML SSO, organization analytics and admin controls. The vendor states SOC 2 and ISO 27001 compliance. The site currently focuses on long-horizon coding-agent workloads, and custom harness integrations require contacting the vendor.
- Per-step model and reasoning-effort recommendation for coding agents
- Works with existing gateways and agent harnesses (not a gateway itself)
- Multi-provider and multi-harness support
- Optional privacy-preserving routing without payload access
- Savings and usage dashboard; continuous learning from feedback
- SOC 2 and ISO 27001; enterprise SSO (SAML) and custom SLAs
As seen on YouTube
Watch the moment: the exact second top creators show it. Videos load only when you press play, via youtube-nocookie.com.
Model routing company that the creator invested in and calls the best on the planet. Mentioned as a way for enterprises to route workloads across models to control costs.
Similar AI tools
Closest neighbours by meaning, not just by category.
Model routing
Research
Technique of sending each AI task to a different model, e.g. a frontier model for planning and a cheaper one for execution.
Strands Decider 2B
Model · Amazon
Small open model from Amazon's Strands agents team, reportedly built for agent decision-making; details unverified.
Paro 26.9
Model · Unbiased
Routing system from Unbiased that dispatches requests across several models with a checking harness; stealth name Union Alpha.
Muse Ember
Model · Meta
Meta's smallest Muse-series language model, positioned as a low-cost option for subtasks and delegation.
Muse Spark 1.2
Model · Meta
Closed-source Meta model for coding and agentic workflows with 1M-token context and multimodal input, offered via API.
Inference.net
API · Inference.net
OpenAI-compatible inference API and gateway for open-weight and closed models, with provider failover and pass-through pricing.