AINewsnow

Bonsai 2 27B: How Ternary Weights and a Hadamard Rotation Compress a 27B Model to 5.9 GB

Bonsai 2 27B: How Ternary Weights and a Hadamard Rotation Compress a 27B Model to 5.9 GB A 27-billion-parameter model that fits in under 6 GB and retains 98.2% of its full-precision benchmark performance sounds like a contradiction. Bonsai 2 27B, released by PrismML on September 17, 2026, makes tha…

Read the full story at DEV Community — Machine Learning ↗

Timeline · 1 report

  1. 2026-09-28 16:07 · DEV Community — Machine Learning
    Bonsai 2 27B: How Ternary Weights and a Hadamard Rotation Compress a 27B Model to 5.9 GB

More stories

  1. NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring — NVIDIA Technical Blog
  2. Launching Meta Enterprise Platform — Meta Newsroom
  3. Heads of OpenAI and Anthropic called to face Senate inquiry after rogue agent incidents — The Guardian AI
  4. Scoop: Anthropic's Dario Amodei to have White House dinner with Trump — Axios AI+
  5. Bill Gates says unchecked AI could ‘cause a billion deaths’ in call for regulation — The Guardian AI
  6. Unsecured OpenAI agents posted 53 user images on the internet without the lab's knowledge — TechCrunch AI
  7. OpenAI’s A.I. Went Rogue and Meddled With U.S. Government Websites — New York Times Technology
  8. Nvidia’s Answer to Rogue Agents Is an Open-Source AI Security System — Wired AI

Get the daily brief of stories like this at 6:30 every morning →