Tether AI Research (QVAC) · On-Device Agent LLM

BotPsy-0.8B

BotPsy is a 0.8B-parameter language model for function calling and agentic workflows, built on Qwen3.5-0.8B for on-device use. Tether reports BFCL v4 rising from 23.4 to 44.0 and τ-bench from 23.2 to 51.5 over its base model, and says it leads all models up to 1.2B parameters tested on six tool-use and agentic benchmarks. It is to be released under Apache 2.0 in bf16, Q8 and Q4 precisions, as safetensors and GGUF files for llama.cpp and the QVAC SDK.

Open the source ↗

Similar Language Models releases

  • Muse Ember

    Meta Superintelligence Labs · Lightweight agent model · Week 47, 2026

  • qwenFS

    vixhal-baraiya (GitHub) · Filesystem-in-Weights LLM · Week 47, 2026

  • Spark-X2.6-1.7B

    XHToken · On-device LLM · Week 45, 2026

  • LFM3

    Liquid AI · Small multimodal LFMs · Week 45, 2026

  • AstaBrief 8B

    Allen Institute for AI (Ai2) · Open Science Report Model · Week 41, 2026

  • Bespoke-Nimble-9B

    Bespoke Labs · Structured Scoring Adapter · Week 39, 2026

Published by Quanta AI Labs, an AI agency in India. Browse every week in AI News Weekly.