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.
Imp brings DSPy's declarative, self-improving language-model programming paradigm to Elixir, leveraging OTP for reliable, concurrent agent execution.
The new model uses Glow-TTS architecture and targets edge devices and CPU inference.
The Emacs package agent-shell, based on the ACP protocol, has been released, enabling control of 19 coding agents through a single UI.
Julia-1 is a 144.3 million parameter multilingual encoder designed for CPU-only execution, capable of classification, ranking, and binary question answering.
Atlas is a Tauri-based desktop application that manages parallel coding agents like Claude Code and Codex by linking commits to session contexts and sharing memory across tools.
The 8B parameter model achieves competitive results with Qwen-Image 20B on specific benchmarks while using a single backbone for both generation and understanding.
The open-source intents-mcp tool exposes 1,304 built-in macOS App Intents to MCP clients like Claude Code without requiring SIP modifications.
Vantage is a free, local-first tool that tracks costs, enforces permissions, and logs sessions for AI coding agents like Claude and OpenAI.
A new pull request introduces tiled matrix multiplication for k-quants in llama.cpp, leveraging VNNI instructions to achieve 3-7x speedups in CPU prompt processing.
PiG offers a static Go binary port of the Pi coding agent, achieving ~21ms startup time compared to Pi's ~300ms median.
The open-source Token-Space Font Compiler generates custom fonts where each token occupies a fixed width, supporting presets for DeepSeek V4.1 Flash, ctok (Claude approximation), and various BPE tokenizers.
Hayder Tirmazi's optimizations for prompt lookup decoding in llama.cpp, combined with a PR by Daniel Lemire, achieve up to 140x faster drafting and up to 2.6x less memory usage.
Mica v0.1 4B is an Apache-2.0 licensed model that outputs calibrated probabilities for agent decisions, achieving 67.0 accuracy on held-out English sets while maintaining a small local footprint.
Ollaya enables local execution of open-source decision models like Laya and Decider, offering 8–10ms latency on NVIDIA GPUs and compatibility with the TypeSafe Python SDK to replace hosted Jev API calls.
The new architecture maintains within 3.4% of native token throughput on NVIDIA H100 GPUs while enabling support for out-of-tree accelerators and older hardware.
Released the open-source tool entail to verify and prevent the issue where vLLM's rope_scaling override bug causes GSM8K scores to drop by up to 28%.
Whiteboard is an open-source VS Code fork that enables human developers and AI coding agents to collaborate on software architecture using a shared canvas.
The new TQ method achieves a mean KLD of 0.0282 and 92.4% top-1 accuracy on Qwen 3.8 27B at 4-bit without requiring calibration data.
The open-source agent generates deterministic MP4s by having the LLM write HTML/CSS/JS code that is rendered via headless Chromium and ffmpeg, rather than using a generative video model.
The model uses a single reinforcement-learning stage without prior domain-specific supervised fine-tuning.
The new Swift 1.5 27B model reduces thinking tokens by 58.5% while improving accuracy, with Flash Next offering a 1.8x speed increase.
The new integration supports a 124B parameter sparse MoE model with 5.5B active parameters and a 256K token context window.
The Qwen3-based CLM supports all of Jev's features and achieved 87.6% performance on Terminal-Bench.
GravityOCR, based on GLM-OCR, increased decoding speed by 1.32x without adding parameters while maintaining an accuracy score of 95.16.