HELAX AI logo HELAX AI Docs
API Playground Get Started
🇱🇰 Built for Sri Lanka

HELAX AI N3

Advanced Sinhala Intelligence for the Next Generation

HELAX AI N3 is an advanced artificial intelligence model developed by HELAX AI with strong Sinhala language capabilities, multilingual understanding, intelligent reasoning, coding assistance, research support, and conversational intelligence.

Production Ready 🇱🇰 Sinhala · English · Multilingual Built by NHB LK
🇱🇰 සිංහල AI
🧠 Reasoning
N3

Model Highlights

A single, focused model — HELAX AI N3 — engineered for Sinhala-first intelligence.

🧠

Advanced Reasoning

Multi-step logic, planning and decision-making across complex problems.

💻

Coding Intelligence

Generate, debug and optimize code across 12+ programming languages.

🇱🇰

Sinhala Language Understanding

Deep comprehension and generation of natural, fluent Sinhala.

🌎

Multilingual Communication

Seamless understanding across Sinhala, English and more.

📚

Knowledge Retrieval

Find and ground answers in your own documents and data.

Fast Responses

Low-latency, production-grade output built for real workloads.

Performance at a glance

How HELAX AI N3 performs across the capabilities that matter most.

🇱🇰 Sinhala Understanding0%
🧠 Reasoning Ability0%
💻 Coding Support0%
🌎 Multilingual Intelligence0%
📚 Research Assistance0%
✨ User Experience0%

Why HELAX AI N3

Built from the ground up for Sri Lankan users, businesses, developers and organizations.

🇱🇰

Optimized for Sinhala Language

Trained with strong Sinhala understanding and natural generation.

🏝

Built for Sri Lankan Users

Designed around local context, businesses and everyday needs.

🔍

Advanced Context Understanding

Tracks long-range context across documents and conversations.

Modern AI Architecture

A clean, efficient pipeline from input to safe, grounded output.

🧩

Developer Friendly

A predictable REST API and playground for fast integration.

🏢

Enterprise Ready

Security, reliability and controls for organizations at scale.

Frequently asked questions

Common questions about HELAX AI, N3 and S Mini.

What is HELAX AI?

HELAX AI is an artificial intelligence platform that builds advanced, Sinhala-first AI models — including HELAX AI N3 and HELAX S Mini — for Sri Lankan users, businesses and developers.

What is HELAX AI N3?

HELAX AI N3 is the flagship HELAX AI model for reasoning, coding, research, automation and multilingual communication, with strong Sinhala language understanding and generation.

What is HELAX S Mini?

HELAX S Mini is a specialized, lightweight Sinhala-only model. It understands many languages as input but always responds in Sinhala, making it ideal for Sinhala conversations, education and local businesses.

Does HELAX AI support the Sinhala language?

Yes. Sinhala is a core focus of HELAX AI. N3 offers strong Sinhala understanding and generation alongside English and multilingual support, while S Mini is dedicated to Sinhala output.

Who develops HELAX AI?

HELAX AI models are developed by NHB LK under the HELAX AI organization, built specifically for Sri Lankan users and the Sinhala language.

Start building with HELAX AI N3

Advanced Sinhala intelligence for Sri Lankan users, businesses and developers.

Overview

HELAX AI N3

A complete profile of HELAX AI N3 — the flagship, Sinhala-trained AI model from HELAX AI, built for Sri Lankan users, businesses, developers and organizations.

Model Name

HELAX AI N3

Developer

NHB LK

Organization

HELAX AI

Model Type

Large Language Model (LLM)

Status

Production Ready

Languages

🇱🇰 Sinhala · English

What is HELAX AI N3?

HELAX AI N3 is an advanced, Sinhala-trained AI model developed by HELAX AI. It is designed with strong Sinhala language understanding and generation, paired with multilingual communication, intelligent reasoning, coding assistance and research support. N3 is built specifically for Sri Lankan users, businesses, developers and organizations — optimized for Sinhala, English and multilingual communication.

🇱🇰

Sinhala-First

Strong native Sinhala understanding and natural, fluent generation.

🌎

Multilingual

Communicates seamlessly across Sinhala, English and beyond.

🛡

Safe by Design

A dedicated safety layer governs outputs for trustworthy use.

Capabilities

Capabilities

HELAX AI N3 delivers Sinhala-first intelligence — spanning language, content, research, coding and business workflows for Sri Lanka.

🇱🇰

Sinhala Language Processing

Understand grammar, nuance and intent in natural Sinhala text.

Sinhala Content Generation

Write fluent, natural Sinhala articles, posts and messages.

Sinhala Question Answering

Answer questions accurately in clear, conversational Sinhala.

🔁

Sinhala Translation

Translate between Sinhala and English with context and accuracy.

🌐

English Communication

Communicate fluently and professionally in English.

🔬

Research Assistance

Summarize, synthesize and cite across large bodies of knowledge.

💻

Coding Support

Write, refactor and explain code in 12+ programming languages.

Business Automation

Automate documents, replies and operational workflows.

🧠

Workflow Intelligence

Plan and drive multi-step tasks across tools and data.

🎓

Educational Support

Explain concepts and tutor learners in Sinhala and English.

🇱🇰 Sinhala-Only Model

HELAX S Mini

A specialized, lightweight Sinhala AI model. It understands many languages — English, Tamil, Japanese, Chinese, Arabic, French, German, Spanish and more — but always responds in Sinhala.

Input · any language

Englishதமிழ்日本語中文العربيةFrançaisDeutschEspañol
HELAX S Mini

Output · always

🇱🇰 සිංහල

Model Card

A dedicated AI assistant for Sinhala-speaking users.

Name

HELAX S Mini

Type

Sinhala-Only Language Model

Developer

NHB LK · HELAX AI

Input

🌎 Multilingual

Output

🇱🇰 Sinhala only

Status

Production Ready

HELAX S Mini is a specialized Sinhala language model developed by HELAX AI. The model can understand multiple languages but always generates responses in Sinhala, making it a dedicated AI assistant for Sinhala-speaking users. It is lightweight, highly optimized and mobile-friendly — with fast inference and low resource usage — built for Sinhala conversations, education, local businesses, government services and Sri Lankan users.

Features

Optimized end-to-end for Sinhala output.

🇱🇰

Sinhala Native Output

Every response is generated in natural, fluent Sinhala.

🌎

Multilingual Input Understanding

Understands prompts in many languages and scripts.

Ultra Fast Responses

A lightweight architecture tuned for low-latency replies.

📱

Mobile Optimized

Low resource usage for on-device and mobile experiences.

🧠

Context Awareness

Tracks conversation context for coherent dialogue.

📚

Educational Support

Explains lessons and answers questions in Sinhala.

💬

Sinhala Conversations

Natural, everyday chat for Sinhala-speaking users.

🏢

Business Assistance

Drafts replies, documents and support in Sinhala.

Performance

Benchmarks across the tasks HELAX S Mini is built for.

🇱🇰 Sinhala Understanding0%
✍ Sinhala Content Generation0%
🔁 Translation to Sinhala0%
💬 Conversation Quality0%
⚡ Response Speed0%
🧠 Context Retention0%

HELAX AI N3 vs HELAX S Mini

Two models, two jobs — pick the right one for your use case.

AttributeHELAX AI N3HELAX S Mini
TypeFlagship LLMSinhala-only model
Output languageSinhala · English · Multilingual🇱🇰 Sinhala only
Input languageMultilingualMultilingual
Best forReasoning, coding, researchSinhala chat, education, business
FootprintFull-capabilityLightweight · mobile-friendly
Inference speedFastUltra fast

Sinhala-first AI for Sri Lanka

Multilingual in, always Sinhala out — HELAX S Mini.

Reasoning Engine

Reasoning Engine

N3 thinks before it answers — decomposing problems, weighing options and planning multi-step solutions.

The reasoning loop

1

Understand

Parse intent & context

2

Decompose

Split into sub-tasks

3

Reason

Evaluate logic paths

4

Decide

Select best path

5

Respond

Generate answer

Hover a stage to highlight it in the reasoning pipeline.

Multi-Step Reasoning

Chains intermediate steps to reach reliable, verifiable conclusions.

Logical Analysis

Applies deductive and inductive logic to structured problems.

Decision Making

Weighs trade-offs and constraints to choose optimal actions.

🜔

Context Understanding

Tracks long-range context across documents and dialogue.

🗺

Planning

Builds ordered plans for agents and multi-stage tasks.

🧩

Problem Solving

Self-corrects and re-plans when intermediate steps fail.

Coding Intelligence

Coding Intelligence

From a single function to a full service, N3 writes production-grade code and explains every line.

Supported languages

PythonJavaScriptTypeScript PHPJavaGo RustC++C# SQLHTMLCSS
app.pyutils.py
Python

Code Generation

Produce complete, runnable implementations from natural language.

🐞

Debugging

Trace failures, explain stack traces and patch root causes.

Optimization

Improve performance, memory and readability of existing code.

📘

Documentation

Generate docstrings, READMEs and API references automatically.

🔌

API Development

Scaffold endpoints, clients and integration code end-to-end.

🗄

Database Design

Model schemas, write queries and design migrations.

Architecture

Architecture

A look inside the HELAX AI N3 pipeline — from raw input to a safe, generated response.

01

User Input

Raw prompt & context enter the pipeline.

02

Tokenizer

Text is split into model tokens.

03

Embedding Layer

Tokens map to high-dimensional vectors.

04

Neural Processing Engine

Transformer layers process representations.

05

Reasoning Engine

Multi-step logic and planning are applied.

06

Memory System

Long-range context is stored and recalled.

07

Knowledge Layer

Grounded facts and retrieval augment output.

08

Safety Layer

Policies and guardrails filter unsafe content.

09

Response Generator

Final tokens are decoded into language.

10

Output Layer

The response is streamed to the user.

API Reference

API Documentation

A clean, predictable REST API. Authenticate with a bearer key and call N3 from any language.

Authentication

All requests are authenticated with an API key sent as a Bearer token in the Authorization header. Keep your key secret — never expose it in client-side code.

bash
curl https://api.helax.ai/v1/chat/completions \
  -H "Authorization: Bearer $HELAX_API_KEY" \
  -H "Content-Type: application/json"

Endpoints

POST/v1/chat/completions

Generate a chat response from N3.

Request
{
  "model": "helax-n3",
  "messages": [
    { "role": "system", "content": "You are HELAX AI N3." },
    { "role": "user", "content": "Explain quantum tunneling." }
  ],
  "temperature": 0.7,
  "max_tokens": 1024
}
Response
{
  "id": "cmpl_n3_8fa12",
  "model": "helax-n3",
  "choices": [
    { "index": 0, "message": {
        "role": "assistant",
        "content": "Quantum tunneling is…" },
      "finish_reason": "stop" }
  ],
  "usage": { "prompt_tokens": 24, "completion_tokens": 180, "total_tokens": 204 }
}
POST/v1/models

List available HELAX AI models.

Response
{
  "data": [
    { "id": "helax-n3", "type": "llm", "status": "production_ready" }
  ]
}
POST/v1/reasoning

Run an explicit multi-step reasoning request.

Request
{
  "model": "helax-n3",
  "problem": "Plan a 3-step migration to a new database.",
  "show_steps": true
}
POST/v1/agents

Create an autonomous agent run.

Request
{
  "model": "helax-n3",
  "goal": "Research competitors and draft a summary.",
  "tools": ["web_search", "file_write"]
}
POST/v1/files

Upload a file for grounding and retrieval.

bash
curl https://api.helax.ai/v1/files \
  -H "Authorization: Bearer $HELAX_API_KEY" \
  -F purpose="knowledge" \
  -F file="@handbook.pdf"

Error codes

CodeMeaning
400Invalid request payload
401Missing or invalid API key
403Insufficient permissions
404Resource not found
429Rate limit exceeded
500Internal server error

Rate limits

TierRequests / minTokens / min
Free2040K
Pro5001M
EnterpriseCustomCustom
Playground

Playground

Prototype prompts against HELAX AI N3. This is a front-end demo — responses are simulated locally.

Research

Research

Technical insight into how HELAX AI N3 is designed, evaluated and kept safe.

📄

HELAX AI N3 Technical Report

A full overview of model design, training methodology and intended use.

Reasoning Framework

How N3 structures multi-step thought, planning and self-correction.

Architecture Overview

The end-to-end pipeline from tokenizer to safe output.

📊

Performance Analysis

Benchmarks across reasoning, coding and multilingual tasks.

🛡

Safety Framework

Alignment, red-teaming and guardrail methodology.

🗓

Research Timeline

Milestones from early prototypes to the stable N3 release.

Security

Security

Enterprise-grade protection for your data, prompts and infrastructure.

🔒

Data Protection

Encryption in transit and at rest with strict data isolation.

🛡

Prompt Security

Defenses against prompt injection and unsafe tool use.

🕶

Privacy Controls

Granular retention settings and zero-training data options.

🔑

Access Management

Role-based access, scoped API keys and audit logging.

Safety Systems

Layered guardrails monitor and filter every response.

🏢

Enterprise Security

SSO, compliance controls and dedicated deployments.

Changelog

Changelog

The journey of HELAX AI N3 from first prototype to a stable production release.

Stable

HELAX AI N3 Stable

Production-ready release with the full reasoning engine, coding intelligence, long-context memory and the complete REST API.

Beta

HELAX AI N3 Beta

Expanded language support, improved planning and the public Playground. Hardened safety layer and rate limiting.

Alpha

HELAX AI N3 Alpha

First internal release introducing the core neural processing and reasoning engine for early evaluation.