Sential Innovations

Building the models and tools that make AI understandable.

Sential Innovations, known simply as Sential, is an artificial intelligence company. We build AI models first, alongside software, hardware, robotics, and the research that makes all of it safer and more useful. Each product serves a clear purpose, from personality AI to agentic AI.

See our research → Kova: coming soon

Introducing the Deriva Series

In development

What we are working on

A family of model series, each built for a distinct job. We show their status honestly: Deriva is in active development, Hikaru and PICO continue at a reduced pace while we focus elsewhere, and the rest remain concepts. See why we reordered priorities.

Language models

In development Deriva Series

Deriva

A general-purpose model with vision. Capable across a wide range of tasks, including reasonable competence at code, but with no single specialty.

Slowed Hikaru Series

Hikaru

Our testbed for Explanatory Reinforcement Learning. Still moving, just slower, while our development effort sits with Deriva first.

Concept Arcus Series

Arcus

A large model built specifically for software. Notes and plans only; no development has begun. The aim is depth in code, not breadth.

Concept Hirose Series

Hirose

A powerful creative generalist, built to help make things across writing and art, but never to replace the people who do that work.

Concept OBIKI Series

OBIKI

Our flagship model: a massive, fully custom system we intend to be genuinely frontier, aimed at coding, bioscience, and cybersecurity.

Concept Yuki Series

Yuki

The smallest, most efficient model we can build, made to punch far beyond its weight. Yuki means snow in Japanese.

Multimodal & robotics

Concept Nero Series

Nero

Built for robotics and wearables. Combines vision, voice, and language into a single model that understands the physical world.

Speech

Slowed PICO Series

PICO

A speech-to-speech model, voice in and voice out with no text in between. Development continues at a reduced pace.

Image generation

Concept Photon Series

Photon

Our image-generation model. Photon-1 turns written prompts into images. A concept for now.

View all nine model series →

Products

Kova: a home agentics platform.

Kova lets you talk to and act through our models around your home: automations, control, and conversation in one place. At launch, and for some time after, Kova runs Sential Innovations models only.

Kova is not yet available. Follow us on social media or check back here for launch updates.

More about Kova →

Placeholder: To Be Updated

Diagram of the 11-point ERL reward scale — ERL reward scale

Research

Reinforcement learning that explains itself.

Standard reinforcement learning gives a model a binary signal: good or bad. Our method, Explanatory Reinforcement Learning (ERL), replaces that with an eleven-point scale and a grader that explains its reasoning.

The result is a model that learns why it was rewarded, not just that it was.

Read the ERL overview →

News

Latest

Sep 4, 2026 Deriva-1, Update 2: rebuilding around the data Two weeks into Deriva-0.4 and 0.5: an overthinking problem the base model won't give up, a wider and more diverse training mix, the removal of OMEGA reasoning, and an early look at Deriva-0.6. Models Aug 28, 2026 Hikaru-1, Update 2: thirty runs of ERL Three months and more than thirty ERL runs: grader poisoning, thin episode data, and the rethink that moved explanations out of the weights and into retries. All internal, nothing released. Models Aug 28, 2026 Verplex is cancelled We are stopping work on the Verplex smart glasses and moving that time into our AI models and research. A deliberate scope decision, not a door closed for good. Community

All updates →

Explore