Chinese AI startup Z.ai releases GLM-4.7, an open-weight model that Z.ai says delivers significant improvements in coding performance compared to GLM-4.6
like 210 — Z.ai 6.24k — Text Generation Transformers Safetensors English Chinese glm4_moe conversational eWeek : Chinese AI Startup Z.ai Takes On OpenAI Via Cheaper Prices Vincent Chow / South Chi...
OpenAI releases GPT‑5.2-Codex, with improvements on long-horizon work through context compaction, stronger performance on large code changes, and more
What excites me most is the leverage this gives developers. … Frederic Lardinois : At this point, I'd almost be disappointed if nobody releases a new model on Christmas Day. The current release cycle...
Allen Institute for AI launches Bolmo 7B and Bolmo 1B, claiming they are “the first fully open byte-level language models”, built on its Olmo 3 models
and every token gets the same compute, regardless of complexity. Benjamin Minixhofer / @bminixhofer : There are also some things Bolmo lets us do which we just can't do using subword-level LMs. For ex...
Zoom says its “federated AI” model, combining its SLM with open- and closed-source models, got 48.1% on Humanity's Last Exam vs. 45.8% for Gemini 3 Pro w/ tools
outperforming other AI models with a 2.3% jump over the previous SOTA. ✨ HLE is one of the most rigorous tests in AI, built to measure real expert-level knowledge and deep reasoning across [image] Xue...
OpenAI says GPT‑5.2 Thinking beats or ties industry professionals on 70.9% of GDPval knowledge work tasks, delivering outputs at >11x the speed and <1% the cost
OpenAI eyes January exit from “code red” John Werner / Forbes : The Wonder And The Promise Of GPT 5.2 Is Here Benj Edwards / Ars Technica : OpenAI releases GPT-5.2 after “code red” Google threat alert...
Google says Gemini 3 Pro sets new vision AI benchmark records, including in complex visual reasoning, beating Claude Opus 4.5 and GPT-5.1 in some categories
Raising Concerns for Real-World Use Will McCurdy / PCMag : ChatGPT Overtakes Amazon, X, Reddit, WhatsApp, and Wikipedia in Visitors X: Demis Hassabis / @demishassabis : Gemini has always had exception...
Google launches Gemini 3 Pro Image, aka Nano Banana Pro, with more control, improved text rendering, and enhanced world knowledge, for free in the Gemini app
except when it gaslit me Ryan Whitwam / Ars Technica : Google's new Nano Banana Pro uses Gemini 3 power to generate more realistic AI images Robert Hart / The Verge : Google's new AI image creator too...
Cognition releases SWE-1.5, a new coding model in Windsurf, saying it partnered with Cerebras to serve SWE-1.5 at speeds up to 13x faster than Claude Sonnet 4.5
lots of new paradigms/UX to figure out. @cognition : Today we're releasing SWE-1.5, our fast agent model. It achieves near-SOTA coding performance while setting a new standard for speed. Now available...
SemiAnalysis launches InferenceMAX, an open-source benchmark that automatically tracks LLM inference performance across AI models and frameworks every night
vendor-neutral suite runs nightly and tracks performance changes over time Tae Kim / Barron's Online : Nvidia Touts Software Advantage in Beating Rivals Like AMD Dion Harris / NVIDIA : NVIDIA Blackwel...
Google releases the Gemini 2.5 Computer Use model, built on Gemini 2.5 Pro's capabilities to power agents that can interact with UIs, in preview via the API
Google released a new Gemini 2.5 Computer Use model today, specially designed … Carl Franzen / VentureBeat : Google's AI can now surf the web for you, click on buttons, and fill out forms with Gemini ...