The models were trained from scratch on 15 trillion tokens and can toggle thinking on and off. The speech model transcribes three hours of audio in one second.
The model nobody could place a few days ago turns out to be Chinese lab Z.ai's, confirming early guesses. The weights ship Wednesday.
At the Ai4 conference three leading researchers argued that AI should stay open. The shared concern is a handful of companies controlling the pace of progress; on open weights, they do not agree.
Anthropic chief executive Dario Amodei set out his view on AI regulation, arguing open weights are not sufficient on their own and pointing to a crisis of trust in the industry.
The US AI framework currently covers only closed models. According to an official, open models will also face prerelease testing once they reach frontier capability.
GLM-5.3 reached the frontier with a third of Kimi K3's parameters. The analysis argues the reason is not distillation but post-training depth and a decade of accumulation.
Writer has launched Palmyra X6, built on the open source GLM-5.2. Its own research says harness efficiency cuts costs more reliably than choosing a different model.
Alibaba's Qwen team has published the 27-billion-parameter multimodal Qwen3.8-27B under the Apache 2.0 licence. The model handles a 262,000-token context natively.