Onton, a San Francisco-based search and discovery company, released Ontology 1, a neurosymbolic model for complex, conversational, multimodal product search, which achieved a mean precision@10 of 0.63…
A new tutorial from the open-source GeoAI project demonstrates a complete workflow for extracting building footprints from NAIP aerial imagery using U-Net with a ResNet-34 encoder, Grounding DINO, SAM…
Thinking Machines Lab released Inkling-Small, an open weights Mixture-of-Experts model with 276B total and 12B active parameters, about a quarter the size of its 975B-parameter Inkling model. The mode…
NVIDIA's NeMo team released Molt, a PyTorch-native agentic reinforcement learning framework designed to be compact enough for researchers and AI coding assistants to grasp, with roughly 8.6K lines of …
AMD released Instella-MoE-16B-A3B, a fully open Mixture-of-Experts language model with 16B total parameters and 2.8B active per token, trained from scratch on Instinct MI300X and MI325X GPUs. The mode…
NVIDIA's Transformer Engine accelerates transformer training on Ampere and newer GPUs by fusing kernels and using BF16/FP8 computation, with FP8 support limited to compute capability 8.9 or higher. Th…
Supabase has open sourced Supabase Evals, a benchmark and framework for testing AI coding agents on real Supabase tasks, scoring agents such as Claude Code, Codex, and OpenCode on building schemas, de…
DeepSeek released DeepSeek-V4-Flash-0731 on Hugging Face and moved its V4-Flash API to public beta on July 31, 2026, claiming major gains in agentic and coding benchmarks from re-post-training rather …
LingBot-Map, an open-source streaming 3D reconstruction pipeline, enables GPU-aware inference and point cloud export through a tutorial that auto-tunes frame limits, camera iterations, scale frames, a…
JetBrains Research has open-sourced KotlinLLM, an IntelliJ IDEA plugin prototype under the Apache License 2.0, which uses Smart macros (asLlm and mockLlm) to generate Kotlin source code at runtime and…
Nous Research released Hermes Agent support for Buzz, Block's open source Nostr-based workspace for humans and AI agents, offering three integration paths: Buzz Desktop managed runtime, relay bridge, …
PolyAI released Dialog-RSN-1, an audio-native dialog model that processes caller audio directly, integrating turn-taking, speech recognition, function calling, and response generation into a single mo…
Omnigent, an open-source framework, enables a policy-governed multi-agent financial research workflow that retrieves live USD-to-EUR exchange rates from the Frankfurter API, delegates drafting to a te…
Google DeepMind released Gemini Robotics 2, a set of three physical AI models for whole-body control, dexterous manipulation, and multi-robot collaboration, moving beyond table-top tasks to full human…
Tencent has open-sourced AngelSpec, a torch-native framework for training speculative-decoding draft models across six architectures, introducing DFly—a block-diffusion drafter with hybrid target cond…
Marktechpost AI Media Inc released Token Saver, an open-source Model Context Protocol extension for Claude Desktop (MIT licensed, v1.0) that uses local hybrid RAG to cut PDF token costs by 92% to 99%.…
Moonshot AI open-sourced MoonEP, an Expert Parallelism communication library for distributed Mixture-of-Experts workloads, under an MIT license. The library guarantees every rank receives exactly S × …
Prompt engineering, loop engineering, and graph engineering are three distinct units of control in AI systems, not competing techniques, according to a Marktechpost analysis. Prompt engineering contro…
Liquid AI released two open-weight bidirectional encoders, LFM2.5-Encoder-230M and LFM2.5-Encoder-350M, built on the LFM2 hybrid backbone with an 8,192-token context. The 350M model achieves a 17-task…
Moonshot AI's Kimi CLI can be configured as a fully non-interactive AI coding agent through a Python wrapper that executes headless commands, enabling automated code inspection, modification, testing,…