--- title: "Tether Releases QVAC BotPsy-0.8B Open-Weight AI Model for Local Devices" description: "The 0.8B-parameter agentic model doubles tool-use performance over its base architecture, outscoring systems up to 2.5 times its size on key benchmarks." url: https://basisdesk.news/news/tether-releases-qvac-botpsy-open-weight-ai-agent published: 2026-11-11T19:46:17.394Z modified: 2026-11-11T19:46:17.394Z section: AI × Crypto author: Basis Desk Newsroom (AI-generated, source-verified) sentiment: neutral tickers: [USDT] tags: [tether, qvac, botpsy, open-source, machine-learning] license: Quote with attribution to Basis Desk (basisdesk.news). Not financial advice. --- # Tether Releases QVAC BotPsy-0.8B Open-Weight AI Model for Local Devices The 0.8B-parameter agentic model doubles tool-use performance over its base architecture, outscoring systems up to 2.5 times its size on key benchmarks. ## At a glance - **What happened:** Tether AI Research launched QVAC BotPsy-0.8B, an open-weight AI agent model optimized for tool use and local execution on personal hardware. - **Why it matters:** The model demonstrates that compact 0.8B architectures can outperform models 2.5 times larger in agentic tasks, enabling private automated workflows without centralized cloud servers. - **Who is affected:** AI developers, mobile device users, and engineers deploying local agentic systems. - **What's next:** Integration of the model across the QVAC ecosystem alongside VisionPsy, MedPsy, and P2P offloading workflows. - **Primary source:** [Tether’s 800M AI Agent Beats Models 2.5x Its Size on Everyday Phone Tasks - Tether.io](https://tether.io/news/tethers-800m-ai-agent-beats-models-2-5x-its-size-on-everyday-phone-tasks) — tether.io ## Key points - Tether AI Research launched QVAC BotPsy-0.8B, an open-weight agentic model designed for local on-device task automation. - BotPsy outperformed Qwen3.5-2B-Think on five of six agentic benchmarks despite having 60% fewer parameters. - The model scored 46.1 on BFCL multi-turn evaluations and 86.1 on tool hallucination prevention under Apache 2.0 licensing. ## Editorial remark - **Context:** Tether has steadily broadened its capital allocation beyond its core stablecoin operations, channeling operational resources into peer-to-peer networks and open-weight artificial intelligence through its QVAC research arm. - **Impact:** Open-source developers and local hardware users gain an optimized sub-1B agentic model that executes system tool calls and function interactions without routing private application data to third-party corporate servers. - **Watch:** Developer adoption metrics across Hugging Face repositories, integration into the QVAC SDK, and performance benchmarks for the planned P2P node-offloading framework. Tether AI Research launched QVAC BotPsy on Nov. 11, 2026, releasing an open-weight 0.8-billion-parameter artificial intelligence model designed to execute multi-step tool calls on consumer hardware [1]. Built atop Alibaba's Qwen3.5-0.8B architecture, the compact model targets local execution to perform routine mobile assistant tasks, including scheduling, messaging, and system management, without transferring personal data to cloud environments [1][2]. The launch coincides with Tether's broader push into decentralized machine learning infrastructure, which previously saw [Tether Capital Backs AIsa to Expand Agentic AI Payments Infrastructure](https://basisdesk.news/news/tether-capital-invests-aisa-ai-payments-infrastructure). Evaluation benchmarks published alongside the open-source weights show BotPsy-0.8B roughly doubled the agentic capacity of its base model across core industry tests [1]. On the BFCL v4 evaluation, the model scored 44.0 compared with 23.4 for Qwen3.5-0.8B-Think, while reaching 58.7 on ACEBench and 64.6 on ToolSandbox [1]. In five out of six evaluated benchmarks—including BFCL v4, ACEBench, τ-bench, BrowseComp-Plus, and ToolSandbox—the 0.8B model exceeded the accuracy of the 2.5-times-larger Qwen3.5-2B-Think, trailing only on Claw-Eval with a score of 46.5 against 48.9 [1][2]. Furthermore, on BFCL v4 multi-turn tasks, BotPsy achieved a score of 46.1 compared with 23.8 for the 2B model, while scoring 86.1 on tool hallucination prevention tests [1][2]. ## Local Architecture and Distributed Strategy According to Tether, BotPsy addresses computational limits on edge devices by specializing in function calling rather than broad factual retrieval, where large parameter counts typically retain advantages [1][2]. In multi-turn customer troubleshooting tests via τ-bench telecom, the model scored 82.5, surpassing both Qwen3.5-2B-Think (69.3) and MiniCPM5-1B-Think (74.0) [1][2]. Tether CEO Paolo Ardoino stated that local deployment reduces the dependency on centralized providers when handling sensitive data such as calendars and emails, granting users direct authority over device permissions [1]. The model weights are published under an Apache 2.0 license in BF16 format, with quantized GGUF variants for deployment in llama.cpp and QVAC tooling [2]. Tether plans to position BotPsy as an operational agent alongside its other domain models, including VisionPsy, TranscribePsy, MedPsy, and TranslatePsy [1]. The project's long-term framework envisions mobile devices delegating complex tasks, such as document analysis and travel planning, over encrypted peer-to-peer networks to home personal computers or decentralized nodes [1]. At the time of writing, $USDT traded at $0.9989 amid a broader cryptocurrency market capitalization of $2.46 trillion. ## Sources 1. [Tether’s 800M AI Agent Beats Models 2.5x Its Size on Everyday Phone Tasks - Tether.io](https://tether.io/news/tethers-800m-ai-agent-beats-models-2-5x-its-size-on-everyday-phone-tasks) — tether.io, 2026-11-11 2. [qvac/BotPsy-0.8B · Hugging Face](https://huggingface.co/qvac/BotPsy-0.8B) — huggingface.co, 2026-11-11 --- Basis Desk Newsroom · AI-generated, source-verified · https://basisdesk.news/about/how-we-use-ai