
Unsloth Shrinks Z.ai's 754B GLM-5.1 Coding Giant to Fit on a Mac
Unsloth ships day zero GGUF quants of Z.ai's 754B agentic coding model, shrinking it from 1.65TB down to 200GB so it can run on a single high-memory workstation.
20 stories

Unsloth ships day zero GGUF quants of Z.ai's 754B agentic coding model, shrinking it from 1.65TB down to 200GB so it can run on a single high-memory workstation.

SEO Machine is a free, open-source Claude Code workspace that replaces $300–$600/month in SEO tooling with a multi-agent content pipeline grounded in real analytics data.

An uncensored, multimodal build of Google's Gemma-4-31B ships with both thinking and non-thinking modes stripped of refusals via weight editing.

An open-source macOS assistant that watches your screen, talks back, and flies a blue pointer to the exact UI element you asked about.
Kentaro Nomura and Takato Horii publish a multi-agent active inference model showing how social norms and creative acts co-construct shared reality
A new proof shows activation steering pushes LLMs into internal states no text prompt can ever reach, with major implications for AI safety evaluations

An open-source MCP server that lets Claude, ChatGPT, and Cursor search, annotate, and edit your Zotero library with semantic vector search built in.

A new INT4 quantization method uses cheap pairwise rotations to close the accuracy gap with FP16 on reasoning models, running at near-AWQ speed.

A new arXiv note argues Anthropic's emotion probes may track situational context, not affect, and could miss the most dangerous alignment failures.

A FAIR/Meta paper shows that reusing old rollouts in a replay buffer cuts LLM RL training compute by up to 40% — without hurting accuracy, and sometimes improving it.

OpenAI shipped a Rust-based renderer and parser for the token protocol that gpt-oss models require to reason, call tools, and stream structured output.

Microsoft's Bing team just dropped a three-model embedding family topping the multilingual MTEB v2 leaderboard with a 32k token context and MIT licensing.

Meta FAIR extends the DINOv2 repository with three specialized backbones for cell microscopy, multi-channel biology imagery, and chest X-ray analysis.

Alibaba's open-source vision-language model turns messy PDFs, chemistry diagrams, flowcharts, and even music sheets into clean structured HTML.

NVIDIA open-sources NeMo Data Designer, a Python framework that generates production-grade synthetic training datasets with built-in validation, field correlations, and LLM-as-a-judge scoring.

A new survey maps every major approach to fabricating network traffic, from Poisson processes to diffusion models and GPTs, and ships a scoring rubric.

AnyCrawl is an MIT-licensed Node.js crawler that converts sites into markdown, runs LLM extraction, and scrapes SERPs with multi-threaded workers.

OpenBMB released a 2B parameter tokenizer-free TTS model that covers 30 languages, designs voices from text descriptions, and streams 48kHz audio in real time.

A minimal plugin turns Claude Code into a coding partner that literally rings your phone when a task finishes, stalls, or needs a decision.

CopilotKit released an open-source framework that lets AI agents stream live HTML, SVG, and 3D scenes into your app instead of just text.