Meta introduces Muse Ember as its smallest, most efficient model, aimed at the repeated, structured work performed by AI agents: tool use, workflow automation, coding subtasks, routing, information extraction, and lengthy document analysis. It supports text, images, audio, and video, and can interpret screenshots, charts, scanned files, recordings, and video without a separate model. Its context window is one million tokens, with outputs up to 512,000 tokens, allowing it to process large codebases or document collections in one call. Meta positions Ember as a high-throughput counterpart to its Muse Wonder model, which handles harder planning while Ember takes on volume.
Benchmark results generally favor Ember over Claude Haiku 5.5 and GPT-6 Luna on agent and coding tasks, though not every test. Ember scores 37.8% on AutomationBench-AA, versus 20.2% and 13.9%, and leads on Agents’ Last Exam, a personal-work composite, Terminal-Bench 4.0, and Coding Agent Index-AA. Haiku scores higher on FrontierCode 2.0-AA, while Ember and Luna tie on GDP.pdf; the Artificial Analysis Index also shows Ember and Haiku tied at their best settings. Meta’s cost comparison says Ember matches Haiku’s top score for roughly one-third the expense and beats Luna’s tested settings. It is available through the Meta Model API for $0.10 per million input tokens and $0.50 per million output tokens. Early enterprise testers emphasize speed, token efficiency, and usefulness for document review, coding, and multi-repository investigations.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Researchers tested 16 LLMs on all 55 TIS-100 puzzles, giving each one attempt and time to refine its code. The models found 133 new performance tradeoffs across 18 puzzles and set three outright cycle records, though human players improved on most solutions within days. (tis-100.wincon.dev)
Australia is considering a new criminal offence for AI companies that fail to take reasonable steps to prevent harm caused by their models. The government is also weighing an AI safety watchdog with power to impose significant fines, following a hack of a Medicare statistics portal involving OpenAI. (abc.net.au)
AI is increasingly becoming a gateway through which people find information. Audiences still resist having AI produce their news. (niemanlab.org)
Nvidia holds a central position in the AI industry, making disruption to the company potentially catastrophic for the wider ecosystem. The race for advanced AI is also intensifying economic and geopolitical conflict, while concentrated corporate power and infrastructure face growing political resistance. (france24.com)
LLM inference can require reading about 26 GB of model weights to generate each token from a 13-billion-parameter model. The explanation shows how model size drives memory bandwidth demands during text generation. (youtube.com)
Deletion Tuning matches full fine-tuning across three LLM post-training methods by zeroing fewer than 0.05% of model weights. Its compact masks also preserve prior-task accuracy better than full fine-tuning, LoRA, and random sparse tuning. (arxiv.org)
Today's best Hacker News stories, summarized and screenshotted, one email a day.