Fine-tuned IBM Granite 3.3 8B for OpenCypher query generation. 85% of Claude 4 performance, 10x faster, 100x less energy. Runs on IBM Power — no data leaves the enterprise.

Distil Labs Enables Rocketgraph’s Private AI on IBM Power with Small Language Models

Your Data Never Leaves Your Enterprise: Faster, Greener, and More Secure AI-Powered Graph Querying

The Partnership

Rocketgraph, IBM, and distil labs deliver performance on par with Large Language Models for graph analytics — without a single byte of customer data ever leaving the enterprise perimeter.

The solution achieves 85% of Claude 4 performance while executing 10x faster and using 100x less energy than cloud-based LLMs. The SLM operates on IBM Power hardware entirely within your infrastructure.

The Privacy Paradox in Enterprise AI

Enterprise teams want AI-powered analytics but can’t risk sending proprietary data to cloud LLMs. The answer: specialized small language models that run fully on-premises.

A Fundamentally Different Approach

Rather than using a general-purpose LLM, we fine-tuned IBM Granite 3.3 8B specifically for Rocketgraph’s OpenCypher query generation:

Training Data

Rocketgraph platform documentation
900+ synthetic schemas from Neo4j datasets
15,000+ training examples
All validated against the Rocketgraph platform

The Technical Innovation

Standard Cypher:

MATCH (d)-[r:EdgeType]->() RETURN d, count(r) AS count

Rocketgraph idiomatic:

MATCH (d) RETURN d, outdegree(d, EdgeType) AS count

The SLM learned Rocketgraph’s proprietary query dialect — something no general-purpose LLM could do without fine-tuning.

What SLMs Mean for Your Enterprise

Complete Data Sovereignty

All processing happens inside your infrastructure
No data transmitted to external APIs
Full compliance with data residency requirements

Deployment Simplicity

Runs on IBM Power hardware
No GPU clusters or cloud dependencies
Standard enterprise infrastructure

The Performance Revolution

Query translation in under 200 milliseconds (vs 2-5 seconds for cloud LLMs)
10x faster response times
100x less energy consumption

Real-World Impact

The collaboration demonstrates that enterprise AI doesn’t require choosing between capability and privacy. Fine-tuned SLMs deliver specialized performance that matches frontier models — on hardware you already own.

The Future of Enterprise AI

Specialized SLMs are leading the way: private, fast, accurate, and deployable on existing infrastructure. The era of sending enterprise data to cloud LLMs for every task is ending.

Distil Labs Enables Rocketgraph's Private AI on IBM Power with Small Language Models