
Packt's TwinLlama-3.1-8B Turns Your Writing Into a Personalized AI for $25
A full production LLM Twin blueprint pairing data ETL, fine-tuning, RAG, and AWS deployment with LLMOps tooling under an MIT license.
20 stories

A full production LLM Twin blueprint pairing data ETL, fine-tuning, RAG, and AWS deployment with LLMOps tooling under an MIT license.

Microsoft's Qlib + RD-Agent stack automates the full quant R&D loop — from hypothesis to backtest — achieving 2x returns with 70% fewer factors for under $10

DeepSeek's 1.6T-parameter open-weight model tops coding benchmarks and cuts long-context inference costs by 90%, under an MIT license.

DeepSeek V4 Flash brings 1M-token context and frontier-adjacent reasoning at $0.14/M input tokens, under MIT license with open weights

DeepSeek releases V4-Flash-Base, a 284B open-weight MoE model with 1M-token context and a new hybrid attention architecture that slashes inference costs

Unsloth packages Alibaba's Qwen3.6-27B into consumer-ready GGUFs, bringing a frontier-level coding model to 16GB GPUs

CUHK researchers release OpenGame, an open-source agent that turns a single text prompt into a fully playable browser game, complete with its own specialized 27B code model and a new benchmark for evaluating interactive software generation.

An open-source Python library adds a deterministic graph layer under AI agents so every decision carries provenance, causal links, and an audit trail.

OpenAI consolidates three of its toughest agent benchmarks into a single MIT-licensed repo covering paper replication, freelance engineering, and smart contract exploitation.

SlugLab open-sourced a software framework that emulates CXL 3.0 memory pools, Type 2 GPU coherence, and disaggregated memory without real hardware.

An open-source clone of Devin's DeepWiki turns any Git repo into an interactive, AI-generated wiki with diagrams, chat, and codemaps.

ZeroClaw ships a single Rust binary that runs an autonomous personal AI agent across 30+ channels, 20+ LLM providers, and even $10 microcontrollers.

A new open source desktop app from Craft aims to unify agent workflows across Claude, Gemini, Codex, and Copilot with a document-first UI.

Yeachan-Heo released OMX, a workflow layer that wraps OpenAI's Codex CLI with agent teams, hooks, git-worktree isolation, and persistent state.

Zhipu's 0.9B parameter open-source OCR model tops OmniDocBench V1.5 at 94.62, beating Gemini 3 Pro and Qwen3-VL-235B on document parsing.

BrowserWing exposes browser actions as MCP commands and Claude Skills, letting agents automate the web without burning tokens on every click.

Alibaba's new 27B dense multimodal model matches Claude 4.5 Opus on Terminal-Bench and beats its own 397B predecessor on coding benchmarks.

Alibaba's new dense 27B multimodal model matches a 397B MoE on coding benchmarks, ships with 262K context and reasoning that persists across turns.

A new experimental ROS 2 library moves GPU tensors between separate processes without CPU round-trips, using CUDA virtual memory descriptors and DLPack.

Astronomer open-sources an MCP server, CLI, and 20+ agent skills that let Claude Code and Cursor write, debug, and deploy Airflow DAGs autonomously