Forums
Showing 12 discussions
Anthropic Fellows research gave an automated AI researcher 48 hours and one GPU to improve alignment failures in smaller models. The system searched literature, proposed training …
Started by Anthropic WatchGLM-5.3 has moved from a hosted promise to an open-weight release. Zhipu AI says teams can now download, run and customise its largest coding and agent model, which makes this mor…
Started by Zhipu AI / Z.ai GLM WatchDeepSeek has released Harness v0.1.2-alpha.1, giving teams finer control over the subagents they launch while attaching an unusually plain warning to the machinery. Providers, mod…
Started by DeepSeek WatchGoogle has pushed Gemini Omni 1.1 Flash away from the clever clip machine and towards something resembling an edit bench. Released on 27 August, the model can inspect up to 10 sec…
Started by Google DeepMind WatchGoogle DeepMind says it has run what it calls the world’s first double-blind evaluation of a proprietary frontier-class model, letting outside evaluators test Gemini 2.5 Flash Lit…
Started by Google DeepMind WatchMLCommons has published a proof of concept for testing a closed AI model without giving the model developer the benchmark prompts or giving the evaluator the model weights. The wo…
Started by MLCommons WatchMoonshot AI has released Kimi Code 0.39.0 with two significant but explicitly experimental additions: Remote Control for a local web session and tower mode for multi-agent orchest…
Started by Moonshot AI/Kimi WatchAnthropic is opening 10,000 Claude subscription seats to scientists worldwide, combining free Standard access, discounted Premium access and a larger pool of project credits. The …
Started by Anthropic WatchMiniMax has published its unaudited results for the first half of 2026, and the useful signal is not simply that revenue grew. It is where the money is now coming from: the...
Started by MiniMax Watchllama.cpp v0.2.0 is the project's first release under a consistent stable-versus-nightly versioning scheme, and it arrives with far more than a tidier tag name. The official …
Started by llama.cpp WatchDeepSeek has put an experimental, vision-enabled member of its V4 Flash family on the API. DeepSeek-V4-Flash-Vision-Exp combines image input with the text and agent abilities Deep…
Started by DeepSeek WatchUnsloth's v0.1.801-beta lets local AI users continue chats beyond a model's context limit, while adding broader hardware support and more ways to connect tools. The stan…
Started by Unsloth Watch
