{"slug": "tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone", "title": "Tether's Tiny AI Agent Beats Models 2.5 Times Its Size at Controlling Your Phone", "summary": "Tether unveiled QVAC BotPsy, an 800-million-parameter AI agent built on Alibaba's Qwen3.5-0.8B architecture, which scored 44.0% on BFCL v4 versus 23.4% for its base model and 37.7% for the larger Qwen3.5-2B, according to Tether's technical documentation. Across six tool-use evaluations, BotPsy posted the highest results on five, averaging 45.0% against 38.8% for the 2-billion-parameter Qwen model, and it scored 58.7% on ACEBench versus 30.1% for its base model. The model is the latest addition to Tether's QVAC ecosystem, which Tether introduced in May 2025 to run AI on user-controlled devices rather than centralized cloud infrastructure.", "body_md": "#### In brief\n\n- Tether's new 800-million-parameter BotPsy AI agent outperformed models 2.5 times larger on five of six tool-use benchmarks in company testing.\n- The model can manage multi-step tasks involving phone settings, messaging and scheduling, potentially without sending personal information to cloud servers.\n- BotPsy expands Tether's QVAC ecosystem, which already includes local AI models for medical applications, translation and speech processing.\n\nTether, the company behind the world's largest stablecoin, USDT, is getting surprisingly good at making tiny AI models punch above their weight.\n\nThe company on Wednesday [unveiled](https://tether.io/news/tethers-800m-ai-agent-beats-models-2-5x-its-size-on-everyday-phone-tasks/) QVAC BotPsy, an 800-million-parameter AI agent designed to carry out everyday tasks on personal devices. Parameters are essentially the knobs that determine how much an AI model can do, and more of them generally translate to improved performance. And for reference, frontier AI models today run in the trillions of parameters.\n\nIn the company's own testing, Tether’s relatively tiny model outperformed competitors with up to 2.5 times as many parameters, suggesting that AI assistants may not need enormous models to get things done.\n\nUnlike traditional chatbots that simply respond to questions, BotPsy is agentic, meaning it’s designed to take action. It can identify which tools an application needs, execute commands in the right order, and adjust when something goes wrong.\n\nFor example, Tether described a benchmark scenario in which an AI assistant needed to send a message but encountered two problems: Wi-Fi was disabled and the inbox was full. BotPsy turned on Wi-Fi, cleared enough storage, and completed the task, while the underlying model repeatedly attempted the same failed operation.\n\nThat's a fairly mundane achievement, but it's precisely the kind of problem that makes AI agents useful outside a chatbot window.\n\nBotPsy is built on Alibaba's Qwen3.5-0.8B architecture, a relatively small language model that Tether trained further to specialize in interacting with software tools. Its 800 million parameters represent the numerical values the model learned during training, and a smaller parameter count generally means less memory is needed to run it.\n\nAccording to Tether's [technical documentation](https://huggingface.co/qvac/BotPsy-0.8B), BotPsy scored 44.0% on BFCL v4, a benchmark that evaluates an AI system's ability to select and execute software functions across different tasks and conversations. Scores reflect a weighted percentage of successful performance across tool-use categories.\n\nThe underlying Qwen3.5-0.8B model scored 23.4%, while the considerably larger Qwen3.5-2B reached 37.7%.\n\nBotPsy also scored 58.7% on ACEBench, which evaluates the accuracy of function calls, including selecting the appropriate tools and arguments. Its base model managed 30.1%, while the 2-billion-parameter competitor scored 55.4%.\n\nThe company tested five models across six tool-use evaluations, with BotPsy posting the highest results on five. Its average score across those tests reached 45.0%, compared with 38.8% for the larger Qwen model.\n\nBotPsy is the latest addition to QVAC, Tether's effort to build an AI ecosystem that runs on devices controlled by users rather than relying entirely on centralized cloud infrastructure.\n\nThe company first [introduced QVAC](https://decrypt.co/319829/meet-qvac-tether-decentralized-ai-god-pocket) in May 2025, positioning the project as an alternative to AI services operated by technology giants. Its longer-term ambitions included autonomous agents capable of interacting with applications and potentially making payments using Bitcoin and USDT.\n\nTether expanded that initiative in October 2025 with a [41-billion-token dataset](https://decrypt.co/345859/tether-decentralized-ai-app-dataset-to-challenge-big-tech-dominance) for training science and engineering models, alongside QVAC Workbench, an application that lets users run AI models locally.\n\nEarlier this year, Tether [released its SDK](https://decrypt.co/363809/stablecoin-giant-tether-launches-toolkit-for-building-local-offline-ai-apps), providing developers with tools to build offline AI applications across mobile and desktop operating systems.\n\nThe company also developed QVAC MedPsy, a medical AI system designed to run on consumer hardware. Its 1.7-billion-parameter version [outperformed](https://decrypt.co/367127/tether-medical-ai-runs-on-phone-outperforms-models-16x) Google's much larger MedGemma-27B on HealthBench Hard, a test of realistic clinical conversations evaluated using physician-developed criteria.\n\nBotPsy is currently available as an open-weight model under the Apache 2.0 license, meaning developers can download, modify, and integrate it into their own applications. Tether has also released compressed versions intended to reduce memory requirements for local deployment.\n\nThe company has not announced a release date for a finished consumer assistant powered by BotPsy. Its published model card also says the training code and training data are not included in the release.\n\n## Daily Debrief Newsletter\n\nStart every day with the top news stories right now, plus original features, a podcast, videos and more.\n\n[Add on Google](https://www.google.com/preferences/source?q=decrypt.co)", "url": "https://wpnews.pro/news/tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone", "canonical_source": "https://decrypt.co/383066/tether-tiny-ai-agent-beats-models-phone", "published_at": "2026-11-11 19:36:02+00:00", "updated_at": "2026-11-11 19:54:07.946383+00:00", "lang": "en", "topics": ["ai-agents", "ai-tools", "large-language-models", "artificial-intelligence"], "entities": ["Tether", "QVAC BotPsy", "Qwen3.5-0.8B", "Qwen3.5-2B", "Alibaba", "QVAC", "USDT"], "also_reported_by": [], "alternates": {"html": "https://wpnews.pro/news/tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone", "markdown": "https://wpnews.pro/news/tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone.md", "text": "https://wpnews.pro/news/tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone.txt", "jsonld": "https://wpnews.pro/news/tether-s-tiny-ai-agent-beats-models-2-5-times-its-size-at-controlling-your-phone.jsonld"}}