AI Briefing#open-source
The latest AI and developer news about #open-source, with the original source and a short summary.
The latest AI and developer news about #open-source, with the original source and a short summary.
The MIT-licensed tool enables unified control of Claude Code, Codex, Cursor, Grok Build, OpenCode, and Google Antigravity via iOS, Android, web, and desktop apps.
The open-source tool supports 2D/3D scene composition and integrates models from Sora, Midjourney, and Grok under MIT and Apache-2.0 licenses.
The open-source package extends durability to Pi, OpenCode, Codex, and Claude Code by logging every step to a durable backend, ensuring agents resume from the exact point of failure without repeating completed model calls.
Tested on Gemini 3.5 Flash and GLM-5, the system improves performance on legal and office benchmarks without changing model weights.
The entire editor, including timeline and effects, is built from extensions that the agent can modify in real time.
The model achieves 11.43% WER on LibriSpeech test-clean with just 1.06 MB of INT8 weights.
Sphinx pauses git pushes to ask developers questions about the logic of AI-generated diffs, blocking pushes until the developer demonstrates understanding.
An open-source decision model trained on 2,676 contrastive data points that flip the correct answer by changing just one fact
Pie is an open-source LLM serving system where user programs (Inferlets) directly control the KV cache and generation loops, reducing latency by up to 15% for agent workloads like ReACT.
The open-source project provides a Kubernetes-native runtime for AI agents, featuring sandboxed execution, state preservation, and resource multiplexing.
The free, open-source skill allows users to contribute unused compute to a swarm solving problems like Millennium Prize Problems and small LLM training.
The 4B parameter model uses a looped architecture with interleaved sliding-window and full attention to scale intelligence on a small footprint.
The 27B instruction-finetuned model is designed to generate falsifiable hypotheses and investigation plans within the DiscoveryRSI agentic harness.
The new model is open-sourced on Hugging Face and ModelScope, while Search & Fetch APIs are now available for $0.75 per 1,000 calls.
The open-source ORL method and dlmrt engine enable diffusion LLMs to reach 7,708 tokens/s on 8x B200 GPUs.
The new LoRA improves coordinate accuracy from 41% to 74% compared to the base model by returning page numbers and bounding boxes for every extracted field.
LLMxRay is a new open-source tool that provides detailed observability for local Ollama models, including prompt caching analysis, token-level streaming visualization, and protocol comparison.
The free open-source plugin queries PyPI, npm, crates.io, GitHub, and arXiv for the last 12 months of data to prevent Claude Code from rebuilding existing solutions.
The draft revision opens for public comments until November 15, 2026, while temporary pre-export registration requirements take effect immediately.
Generation speed improved 4.6x compared to llama.cpp, enabling 125B MoE model execution with 12GB VRAM
The update enables Multi-Token Prediction for the 12B-parameter Mellum2.1 Thinking model, which features 2.5B active parameters and a 131k context window.
The new STQ1_0 format achieves 1.3125 bits per weight, reducing the model's storage footprint from roughly 1.5TB while maintaining performance comparable to IQ1_M.
The dataset covers 22 Indic languages with native and romanized scripts, generated via an open-source NeMo Data Designer pipeline.
The model supports multimodal retrieval across 34 languages for text and 18 for images, with open commercial use rights.