
Microsoft's MAI-Transcribe-1.5 Transcribes an Hour of Audio in 15 Seconds
Microsoft's MAI-Transcribe-1.5 hits #3 on accuracy benchmarks while processing audio more than twice as fast as any comparable model
20 stories

Microsoft's MAI-Transcribe-1.5 hits #3 on accuracy benchmarks while processing audio more than twice as fast as any comparable model
NVIDIA's NemoClaw blueprint pairs Hermes Agent with a sandboxed runtime so enterprise AI agents learn and improve from every team interaction

NVIDIA's Cosmos 3 collapses five separate physical AI models into one open omni-model, cutting training cycles from months to days
GitHub launched a dedicated Copilot desktop app, sandboxes, an SDK, and CLI voice mode to manage multiple parallel coding agents from one place.

Factory Router automatically picks the cheapest model that can handle each task, cutting costs 20-25% while matching frontier performance on enterprise coding benchmarks.
Perplexity Computer will route each task between a local model on your machine and frontier cloud models automatically, keeping private data on-device.

Krea's first fully in-house image model lands at #6 globally, beating models twice its price with a creativity API that treats aesthetics as a first-class parameter
Krea 2 lands at #6 on the Artificial Analysis text-to-image leaderboard, beating every model except those from OpenAI, Google, and NVIDIA, with weights coming soon.

Google DeepMind's Co-Scientist uses a multi-agent 'tournament of ideas' to generate and rank thousands of scientific hypotheses, now open to individual researchers via Gemini for Science.

Microsoft's Build 2026 overhauls Windows 11 for developers with native Linux tooling, built-in container support, AI-assisted terminals, and OS-enforced agent security boundaries
Datalab's new 650M parameter open-source OCR model scores 83.3% on olmOCR-bench, beating every system under 3B params while running on a MacBook.
Microsoft's new compact desktop packs NVIDIA's Blackwell GPU and 128GB of unified memory to run 120B-parameter models locally without touching the cloud
Microsoft unveils Aion 1.0 SLMs, expands Windows AI APIs beyond NPUs to CPUs and GPUs, and ships agent sandboxing at the OS level.

OpenAI's Codex gets Sites, six role-specific plugins, and Annotations — turning the coding agent into a full enterprise work platform for non-developers
Nous Research ships Hermes Desktop in public preview, putting its open-source self-improving agent into a native Mac, Windows, and Linux app under MIT license.
Researchers from MIT and MIT-IBM Watson AI Lab show that making convolution filters input-dependent yields a 1.33x–1.60x compute advantage over standard Transformers at scales up to 2B parameters.

Replit Canvas gets multi-edit, GPT-Image 2, Seedance video generation, and animated SVGs — turning it into a full agentic design-to-code workspace

NVIDIA open-sources cuVSLAM, a CUDA-accelerated visual SLAM library supporting 1–32 cameras with real-time performance on Jetson edge hardware

H Company's Holo 3.1 brings computer-use agents fully on-device, from MacBooks to DGX Spark, with 2x speed gains and a major mobile leap

Warp's SSH extension now brings its full coding suite — file tree, code review, and AI agent — to remote machines over SSH