NVIDIA released Nemotron 3 Diarization, a small open model that tops VoiceArena's leaderboard while running live conversation streams.
The Chinese lab squeezed 600B parameters into a sparse model that activates 27B per token and holds context…
Qwen-Image-2.1 is small enough for a consumer GPU and handles transparent layers natively.
A new startup sells abliterated open models that refuse nothing, and security researchers are split on the risk.
Harvey releases Tenet, a research-preview model post-trained on long-horizon legal work from an open Kimi K3 base.
Speculative-decoding checkpoints from Liquid AI boost LFM2.5 throughput by up to 3.18x without changing outputs.
Google DeepMind says Gemma has hit one billion cumulative downloads with more than 100,000 community variants.
MiniMax released open weights for a music model that generates complete songs of up to five minutes in…
The open-weights edge model adds tool calling and strong screen grounding to the LFM family.
LTX-2.5 brings open-weights video generation to ComfyUI with multishot and audio support.
Writer's Palmyra X6 model cuts agent costs by 52 percent as enterprise token bills climb.
Z.ai's GLM-5.3 skips retraining and scales up post-training, with an unexpectedly strong cyber capability.