ConwayResearch/Underdog-Saluki-27B-1.0
ConwayResearch published the model Underdog-Saluki-27B-1.0 on Hugging Face.
Key points
- Underdog, the on-device assistant from Conway Research, has released Saluki 27B under Apache 2.0.
- Underdog Saluki 27B is a 2-bit GGUF of Qwen3.8-27B that fits in 7.89 GB.
- Underdog tuned the compression to protect tool calling, the skill that turns a chat model into an agent.
- For developers, that means a 27B-class agent model that runs in stock llama.cpp.
Sources (2)
- [1]ConwayResearch/Underdog-Saluki-27B-1.0Hugging Face trending models · Oct 8, 01:47 AM
ConwayResearch published the model Underdog-Saluki-27B-1.0 on Hugging Face.
- [2]Meet the Underdog Saluki 27B: A 2-bit Qwen3.8-27B That Beats the Original at Tool CallingMarkTechPost · Oct 9, 07:23 AM
Underdog, the on-device assistant from Conway Research, has released Saluki 27B under Apache 2.0.
Underdog Saluki 27B is a 2-bit GGUF of Qwen3.8-27B that fits in 7.89 GB.
Extractive summary: sentences quoted from the sources.
Hype check
Independently evaluated2 independent looks since
- Evaluation · Oct 8, 2026Closed-loop evaluation of LLM agents for embedded software development
- Evaluation · Oct 8, 2026Evaluating Local Language Model Agents for Reproducible Data Engineering: An Empirical Software Engineering Study of Mobility Workflows
- Follow-up · Oct 8, 2026Dynamics as Code: On Model Compression via Dynamic System
- Follow-up · Oct 8, 2026When Interfaces Speak: Data-Aware Generative UI Harness for Active Interaction
- Follow-up · Oct 8, 2026Balancing Reference Guidance and Free Generation in Trajectory Rollouts for Reasoning RL
- Follow-up · Oct 8, 2026SFT-as-Context Mitigates Forgetting in Supervised Fine-Tuning
Before this
- Oct 7, 2026nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4-LiLiCorr
- Oct 7, 2026unslothai/unsloth v0.1.904-beta: Train your own Decision model
- Oct 6, 2026nvidia/GLM-5.3-Flash-FP4-Local
- Oct 6, 2026unslothai/unsloth v0.1.903-beta: New Browser + Voice Cloning
- Oct 6, 2026huggingface/transformers v5.19.0: Release v5.19.0
- Oct 6, 2026[AINews] Reflection Beam - 501B-A23B American Open Model
