LiveUpdated 2026-10-30 21:12 UTC

Microsoft trained a language model on 4 TRILLION tokens using just three possible weight…

Name: Microsoft trained a language model on 4 TRILLION tokens using just three possible weight…
Uploaded: 2026-10-29T04:58:51.000Z
Channel: Sandhya
Description: Microsoft trained a language model on 4 TRILLION tokens using just three possible weight values: -1, 0, and +1. And the GitHub repo behind it has crossed 41,000 stars. It's called BitNet, and the engineering is fascinating. Most language models rely on weights stored in formats like FP16 or BF16,…

This is a dev post classified by AGK Intelligence as AI dev tools (a tool drop), kept by the Dev Radar because it carries real work, not commentary.

Microsoft trained a language model on 4 TRILLION tokens using just three possible weight values: -1, 0, and +1. And the GitHub repo behind it has crossed 41,000 stars. It's called BitNet, and the engineering is fascinating. Most language models rely on weights stored in formats like FP16 or BF16, requiring 16 bits per weight. BitNet b1.58 uses ternary weights, meaning each quantized weight has only three possible states. That's where the 1.58-bit name comes from. Three states contain approximately 1.58 bits of information. Microsoft didn't simply take an existing full-precision model and

Posted by Sandhya (1.4k followers) 1 days ago · 446 likes · 59.2k views · view the original post on X. Kept by the Dev Radar as AI dev tools. Tools mentioned: bitnet.

More dev work like this

Rust port of pi, a minimal coding-agent harness: single native binary with full… — @rustaceans_rs
🤯天呐！Pi 和 Claude Code 用户狂喜！工具Token调用狂砍 40%！LiquidAI LFM2.5-24B-A2B + HF 专属 LoRA 适配器来了！ — @NFT_Chen
Mistral Large 4, aka Le Chonk, is now live on llama.cpp (a 1T-parameter model). — @TeksEdge
🎃Dario给中国人民带来的极致恐怖从Claude无情封号开始 — @NFT_Chen
Mia's Qwen3.8-Flash-Next recipe on a DGX Spark: one command, first try, 2 min to the… — @niklaslenz_ai
Meet Audio8-TTS-Realistic-0.6b: a 0.6B text-to-speech model that clones voices from just… — @HuggingModels
We're happy to announce our Ruby sponsor: @Superconductor, the multiplayer AI workspace… — @sfrubyconf
THIS IS ABSOLUTE F*CKING GOLD — @marfinxx

Every post is read and classified by AGK Intelligence: what it is, which market it belongs to, and whether the link is a real tool. 31.2k posts from 5k X accounts over the last 21 days, 8.8k tools, 12 markets. New posts are collected several times an hour and the full ranking is recomputed every hour — last update 2026-10-30 21:12 UTC. Full method.