#openweightmodels

Meta has released the weights for Muse Spark Open, describing it as the most capable open-weight...

TL;DR * Reflection AI announced Beam on October 5, 2026, a 501B-parameter (23B active)...

Meta released Muse Spark Open on Oct 15, 2026: an 816B-parameter open-weight multimodal MoE model built for coding and agentic work. Here's what it me

Nvidia in Early Talks to Acquire or Deepen Investment in Open-Weight AI Startup Reflection...

Qwen 3.8 Max is Alibaba's 2.4-trillion-parameter open-weight flagship that claims to trail only Claude Fable 5. Here is what is verified, what it cost

Which open-weight AI model licences let you run, modify and sell freely? A practical guide to Apache 2.0, MIT, Llama and Gemma terms for self-hosting teams.

Xiaomi's MiMo-V2.6-Pro just topped Artificial Analysis's Intelligence Index among open-weight models, tied with Grok 4.7, at a fraction of the cost of proprietary rivals. Xiaomi also open-sourced the full training report, 7,000+ RL environments, and livestreamed the entire run.

List prices per token rose on 10 of 14 model tiers in a year, while the cheapest model scoring 40 on a benchmark index fell from $10 to $0.20 since April.

IFM released K2 Horizon as six Apache 2.0 models spanning 375 billion down to 0.9 billion parameters that share architecture, vocabulary and training methodology, publishing intermediate checkpoints, data-construction recipes, training code and logs

NVIDIA entered a definitive agreement on September 2, 2026 to acquire Hugging Face for approximately $12.93 billion, with its SEC filing stating the deal is expected to close in the first half of 2027 pending regulatory approval -- meaning the acquis

A study on stealing reasoning traces found that Kimi K3 responds unusually strongly to Claude's decoded reasoning, but the authors state plainly that their results cannot establish memorization or distillation, and that reproducing even 16 tokens ver

TielCoder, a 4-bit re-quantization of the Ornith-1.5 mixture-of-experts model, fits in 22.4 GB and fixed 12 of 25 recent real-world coding issues in its maintainer's tests, the same count as Claude Opus 4.6 at medium effort.