LONDON, October 14, 2026 – As AI is implemented into workflows, the widespread reliance on centralised cloud-hosted LLMs has introduced significant data privacy vulnerabilities, regulatory exposure and cloud data collection risks. This is coupled with standard high-cost LLMs taking seconds to return an output, generating unnecessary text and demanding large computing power.
FLock.io, a developer of privacy-preserving AI infrastructure, has launched THIS/THAT Model: an open-source, lightweight decision model developed on Alibaba Cloud’s Qwen3.5-2B edge-friendly architecture. Designed to quickly handle high-volume classification tasks at a low cost locally on edge devices, the system eliminates the need to send sensitive data through external cloud APIs.
The rise of System One/decision models
A new wave of System One models, also known as decision-models, such as TypeSafe’s Jev, has been disrupting enterprise AI since September. Unlike standard LLMs, which generate open-ended natural language prose, they are for queries that have a predefined set of possible answers, such as yes/no.
The name comes from the psychological framework of “system 1” thinking (fast, automatic) versus “system 2” thinking (slow, effortful). Chatbots such as ChatGPT do the latter.
Jiahao Sun, CEO of FLock.io, said: “For many critical enterprise tasks, generating open-ended text is an unnecessary cost and performance bottleneck. With THIS / THAT Model, we are demonstrating that specialised 2B-parameter models running locally can outperform massive cloud LLMs in speed, accuracy, and cost-efficiency – all while keeping data strictly on-device, minimising the attack surface and reducing compliance risk.”
Edge performance
THIS/THAT has outperformed frontier larger LLMs, despite standing at an estimated <0.04% of the parameter size of massive closed-source cloud models like Claude Opus 5. It delivers local execution in approximately 31.4 milliseconds – roughly 30x to 100x faster than cloud-hosted LLMs, eliminating network latency and public API dependencies. It functions completely offline with zero dependency on external cloud servers, preventing service disruptions during network outages.
The model reached an 87.8% success rate across 1,710 complex decision-making tasks in internal benchmarking, outperforming giant closed-source alternatives. In local-geometry questions, it beat Jev on accuracy.
This specialised model is designed for tasks where quick decisions and data classification matter most – in such cases, the slow, costly natural language generation of chatbots is unnecessary. Three weeks after going live, it has already processed 2.5 million decisions.
THIS / THAT Model’s specs
THIS / THAT Model is small enough to run on standard developer hardware and edge devices, meaning data can stay local and latency can be ultra-low. It was trained on Alibaba Cloud using synthetic training data generated using Qwen Max, and built on Qwen3.5-2B: a tiny, edge-friendly foundation model that can be run locally on laptops and smartphones.
It is lightweight, with only 1.88B parameters (less than an estimated 0.04% the size of Claude Opus 5) and zero generated output tokens. The local inference latency is ultra low, at approximately 31.4 milliseconds, compared to standard LLMs taking seconds. It drew on key findings from FLock.io’s recent research papers, GraphSynth and CircuitSynth, accepted to the prestigious ACL 2026 conference.
Background on Qwen and FLock.io’s partnership
Qwen, a tier 1 LLM developed by Alibaba Cloud, partnered with FLock.io in June 2025. It marked the first time a major Web2 centralised LLM has chosen to deeply integrate with a decentralised training platform.
About FLock.io
FLock.io is an AI research and infrastructure company specialising in enterprise-grade federated learning and privacy-preserving AI solutions (AI Arena, FL Alliance, FOMO and the API Platform). Its decentralized architecture enables organisations to train, optimise and deploy custom AI models on local hardware while maintaining total data privacy, model ownership, and regulatory compliance by design.
Media contact: [email protected]
