Clevi 로고

Clevi Models

Clevi provides various AI models tailored to client needs and supports a strong security on-premise installation option.

Ivy-3-text

Flexible response to various business scenarios such as AS (After Service) requests

Ivy-3-text is a lightweight LLM model designed to perform a wide range of specialized roles even at a small scale. With quantization technology, it can be implemented in edge computing and on-device solutions, making it widely applicable in environments such as drones, autonomous vehicles, intelligent CCTV, smartphones, humanoids, and industrial facilities.

Fast Response and High-Quality Answers

  • Despite its relatively small size, Ivy-3-text delivers high-quality answers quickly through an advanced architecture.
  • This enables immediate information delivery and user support in various environments.

Simultaneous Support for Multiple Users

  • Maintains stable performance even with many concurrent users, providing seamless customer support and information delivery.

Easy Maintenance & Expansion, Maximized Business Efficiency

  • The small model size reduces infrastructure burden and allows for easy expansion or integration as needed.
  • It flexibly meets business requirements of various scales.
model ivy-3-text image

Cip-4-mm

Multimodal non-reasoning model

Cip-4-mm is a multimodal AI model capable of processing various input data such as natural language, images, and audio simultaneously. It specializes in contextual connection and pattern recognition rather than complex logical reasoning, excelling in fast information processing such as search, classification, and summarization.

Broader Input Channel Integration

  • Handles text, images, audio, and more at the same time, enabling flexible use in multimodal environments.

Fast Search & Analysis-Oriented Functions

  • Excels at searching, classifying, and matching large-scale documents and data, rather than complex logical reasoning.

Maximized Efficiency and Scalability

  • With high throughput and a flexible structure, it can be easily expanded to meet various organizational needs and operates stably even in large-scale environments.
model cip-4-mm image

Cip-5-x

Top-performance reasoning model

Cip-5-x is a premium AI model with the highest level of reasoning capabilities, optimized for solving complex problems and step-by-step logical development. It provides precise analysis, planning, and decision-making support, delivering outstanding performance in various high-level tasks.

Step-by-Step Reasoning with Clear Evidence

  • Cip-5-x clearly explains the basis of conclusions through step-by-step reasoning, even for ambiguous questions.
  • When used for extensive research and verification, it makes it easy to identify learning and improvement points.

Outstanding Performance in Science & Mathematics

  • Understands and applies complex formulas and theories to provide accurate calculations and interpretations.
  • Delivers excellent results even in high-level tasks such as data analysis and numerical modeling.

High Reliability and Reproducibility

  • The logical process of completed answers can be verified, increasing reliability, and the same process can be reproduced for collaborative or joint research.

Scalability for Research & Development

  • Easily scalable to various research and development environments, making it the optimal solution for projects requiring high expertise and reliability.
model cip-5-x image

Cip-5-agent

Agent-purpose reasoning model

Cip-5-agent is a next-generation agent-type AI model with the ability to understand situations and plan, execute, and adjust to achieve goals. It interprets user instructions and automatically processes step-by-step tasks, creating actionable results by integrating with various tools and systems.

Goal-Oriented Reasoning

  • Goes beyond simple Q&A to understand user goals and establish plans to achieve them.
  • Automatically executes step-by-step actions for efficient handling of complex tasks.

Utilization of Various Tools

  • Integrates with external APIs, data, and systems to deliver actionable results.

Agentic AI

  • Can be directly applied to various purpose-based agent designs such as coding agents and workflow automation agents.
  • Intelligently coordinates and executes complex workflows, providing an optimized solution for business automation and productivity improvement.
model cip-5-agent image

Cip-5-vision

Agent-purpose reasoning vision model

Cip-5-vision is a next-generation intelligent vision AI model specialized in understanding and reasoning with video and image-based data. It interprets visual context to support decision-making connected to real-world environments and can precisely analyze complex visual information.

Visual Information Understanding

  • Accurately recognizes and infers the meaning of objects, relationships, and situations within images and videos.

Precise Analysis and Interpretation

  • Understands and applies complex formulas and theories to provide accurate calculations and interpretations, delivering excellent results even in high-level tasks such as data analysis and numerical modeling.

High Reliability and Reproducibility

  • The logical process of completed answers can be verified, increasing reliability, and the same process can be reproduced for collaborative or joint research.

Scalability for Research & Development

  • Easily scalable to various research and development environments, making it the optimal solution for projects requiring high expertise and reliability.
model cip-5-vision image
Model | Clevi