Abacus AI's Bindu Reddy announces Smaug Nemotron 3.5, an agentic coding fine tune of NVIDIA's Nemotron 3.5 Super VL that she calls the best US open model for a single machine
🚨 Announcing Smaug-Nemotron-3.5 - The Best US Open-Source LLM That Can Run On Your Desktop We are super excited to push the boundaries of US open-source. Smaug-Nemotron improves on Nvidia's latest model that literally runs on a single machine; all you need is one DGX Spark, one H200, or one GB200! The Smaug variant is optimized for agentic coding, improves Deep-SWE, and enables local agents. It's the best US agentic open-source model that can run on a single machine. Weights are now available on HuggingFace
Context
It shows open US models being tuned for agent coding on hardware one team can own. The gains reported on the model card are small, a few points on coding benchmarks, and the best model claim is the company's own.
Bindu Reddy is the CEO of Abacus AI, which has released earlier Smaug models. According to the Hugging Face model card, the model is a supervised fine tune of NVIDIA Nemotron 3.5 Super VL, a 120B parameter model with 12B active, and a 4 bit NVFP4 version fits on one H200, two H100s or one GB200. The post was published at 12:00 AM Eastern on October 23.
Deep dive
Bindu Reddy posted just after midnight Eastern on October 23 announcing Smaug Nemotron 3.5, which she called the best US open source LLM that can run on a desktop. That ranking is the company's claim and has not been checked by an independent leaderboard.
The Hugging Face model card describes the release as an agentic fine tune of NVIDIA Nemotron 3.5 Super VL 120B A12B, a hybrid Mamba 2, mixture of experts and attention model with vision. Abacus AI says it changed behaviour only: the architecture, tokenizer and vision encoder are unchanged, so any stack that serves the base model can serve this one. Training used filtered multi turn, tool using coding trajectories.
The reported gains are modest. For the full precision version the card lists plus 8.4 on Terminal Bench 2.1, plus 2.7 on DeepSWE and plus 1.2 on SWE bench Verified, with reasoning and long context scores held near the base. The 4 bit NVFP4 version, shown in the post's chart, lists smaller gains, such as plus 3.0 on DeepSWE and plus 0.7 on SWE bench Verified, with IFBench slightly lower. All of these numbers come from Abacus AI's own evaluations.
The post mentions running on one DGX Spark, one H200 or one GB200; the card says the NVFP4 weights take about 75 GB. The model inherits NVIDIA's OpenMDW 1.1 license.
What to expect next
Independent results are the next check, for example from Artificial Analysis or community runs on SWE bench and Terminal Bench, since the card's gains are small enough to fall within run to run noise. Watch whether NVIDIA's own next Nemotron update narrows the gap, and whether downloads pick up on Hugging Face. Developers who already serve Nemotron 3.5 Super VL can test it as a drop in swap.