Stellr Intelligence Layer

Models

Medium Language Models for compressed context, persona-native behavior, and lower token usage - without oversized compute or cost.

No per-token billing 256K context ceiling Plug-n-play architecture
In Training

Cintex Beta 1

Medium-scale model for efficient reasoning and persona-native workflows.

Context 256,000 Entry Plus Plan

Input

Text

Output

Text
+
+
+
+
01 // Introduction

Built for efficient reasoning.

Stellr is preparing a family of Medium Language Models around a custom plug-n-play transformer architecture. The goal is not to chase unnecessary scale, but to ship focused, adaptable models you can compose, tune, and deploy around specific workflows.

All models prioritize context switching, persona-native behavior, dense context maps, and lower token usage - so you expand capability without intensive compute or cost.

// Context switching

Heavy yet stable switching between roles, tools, tasks, and knowledge domains.

// Persona-native

Personas shape behavior, priorities, tone, and decisions at the architecture level.

// Compressed context

Long interactions become dense maps that preserve intent, constraints, and memory.

// Reduced token usage

Optimized I/O and internal representation so meaningful work costs fewer tokens.

+
+
+
+
02 // Architecture

Modular blocks. No stack rebuilds.

01

Shared communication

Offload inference batches to approved devices via Exegesis Websocket - steady links and cache state.

02

Batched training updates

Fine-tune personas in batches for subtle, steady gains in ability and knowledge.

03

Small-n-Mid scale

Specialized personas load on demand - fast context switches with capable tools.

Architecture Plug-n-Play Scale Small · Mid Token pricing None
03 // Plans

Plans

We don’t price per-1M-tokens. Pick a request frequency tier and ship. All plans share the same context ceiling.

Tier 01

Plus

$10 /mo
  • 500 RPM
  • No input/output token costs
  • 256,000 context
Coming soon
Popular
Tier 02

Pro

$20 /mo
  • 2,500 RPM
  • No input/output token costs
  • 256,000 context
Coming soon
Tier 03

Ultra

$35 /mo
  • 10,000 RPM
  • No input/output token costs
  • 256,000 context
Coming soon

All plans · No per-token billing · Same 256K context ceiling