Run Medical Report Inference Locally on Mac Studio
MaziyarPanahi · x · 2026-07-17
The author tested Thinking Machines' Inkling medical capability on a Mac Studio: first, it "looks" at a photo of a doctor's consultation/lab report, extracts 23 metrics from the image, then answers questions based on that data.
Results showed it identified hyperkalemia with renal failure as the most urgent issue; gave a clear "no" to whether a contrast CT could be done tomorrow while on metformin, explaining the risk of lactic acidosis from eGFR + contrast, and automatically generated a handoff summary. The author notes this is a 975B parameter, 1-bit version model, with image input and inference done locally using Unsloth, llama.cpp, and Hugging Face.
Related event: Multimodal Model Inkling Tested Locally for Medical Report Reading(3 posts)→
More from Models
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- OpenAI’s Codex + GPT-5.6 Sol hits 99% recall in Project APE verification tests — soumitrashukla9 · 2026-07-22
- OpenAI rolls out voice in GPT-Live, but the UI obscures search and reasoning — Graham_dePenros · 2026-07-22
- Moonshot’s Kimi K3 sets a new open-weights ECI record at 156 — scaling01 · 2026-07-22
- Nanbeige4.2-3B launches as a 3B Looped Transformer model that beats larger baselines — Wooden-Deer-1276 · 2026-07-22
- A post says six companies now beat Google’s best LLM, including two open-source models — soham_btw · 2026-07-22