FLock's THIS/THAT Model Achieves 87.8% Accuracy, Outperforming GPT-5.6 and Claude Opus 5

iconKuCoinFlash
Share
AI summary iconSummary
FLock's THIS/THAT model achieved 87.8% accuracy across 1,710 decision tests, outperforming GPT-5.6 and Claude Opus 5. Built on Qwen3.5 2B and Qwen Max data, it leverages FLock’s GraphSynth and CircuitSynth research. The model features 18.8 billion parameters, 31.4ms latency, and runs locally. On-chain data analysis demonstrates strong performance in fear and greed index scenarios.
ME AI message: The open-source decision model THIS / THAT, developed by the decentralized AI platform FLock.io, has received official recognition and promotional support from Alibaba’s Qwen. Built on Qwen3.5 2B, the model leverages Qwen Max to generate training data and integrates two research breakthroughs—GraphSynth and CircuitSynth—published by the FLock Research Team at the international academic conference ACL 2026 in natural language processing. Training was completed on the Alibaba Cloud platform. Previously, FLock has established long-term technical and ecosystem partnerships with Qwen and Alibaba Cloud, and jointly hosted an AI Hackathon with Qwen at universities in South Korea, continuously advancing AI innovation and developer ecosystem growth.
According to test results released by the team, the THIS / THAT Model 1.2 achieved an 87.8% success rate across 1,710 complex decision-making tests, a significant improvement from the 40.6% of version 1.0, surpassing GPT-5.6 (81.6%) and Claude Opus 5 (83.4%). With only 1.88 billion parameters and no need to generate output tokens, the model has a single-decision latency of approximately 31.4 milliseconds and can run on local devices, making it suitable for AI Agents, enterprise AI, and automated workflows.
FLock.io stated that this model achievement further demonstrates FLock’s technical capabilities in original AI research, data synthesis, and high-performance specialized model training, and represents another significant outcome of FLock’s long-term technical and ecosystem collaboration with Qwen and Alibaba Cloud. In the future, both parties will continue exploring the application of efficient, privacy-preserving AI models in more real-world scenarios. (Source: Foresight News)
Source:Show original
Disclaimer: The information on this page may have been obtained from third parties and does not necessarily reflect the views or opinions of KuCoin. This content is provided for general informational purposes only, without any representation or warranty of any kind, nor shall it be construed as financial or investment advice. KuCoin shall not be liable for any errors or omissions, or for any outcomes resulting from the use of this information. Investments in digital assets can be risky. Please carefully evaluate the risks of a product and your risk tolerance based on your own financial circumstances. For more information, please refer to our Terms of Use and Risk Disclosure.