The research report shows a Reachy Mini robot running on a Jetson Orin Nano 8GB. In that demo, Cosmos 3 Edge handles vision ...

Anthropic sandbox-runtime gives AI agents OS-level sandboxes without containers; Anthropic reports 84% fewer Claude Code ...

Moonshot AI’s Kimi K3.1 Pro is a 2.8T-parameter open-weight model with 1M context, scoring 65.7% on Terminal-Bench 4.0.

CADRE is a family of video autoencoders from Kyutai built for action-conditioned world models. A world model generates in the autoencoder’s latent space. So the autoencoder’s compression sets both ...

TinyFish's agentic search architecture runs 7 phases; Search and Fetch scored 79.0% grounded completion at $0.050 per task.

Liquid AI just announced Liquid RIPL, short for Real-time Inference Pipeline. RIPL is a low-latency engine that serves Liquid ...

Gradium Phonon is a 31M-parameter on-device TTS model under Apache 2.0, reporting 1.70% WER on hard English sentences.

Perplexity’s JOLT trains one model as teacher and student, matching GRPO’s MATH accuracy with 6.5x fewer training tokens.

Recursive Flint is a 40B MoE local coding model with 2.5B active parameters, scoring 74.2% on SWE-bench Verified.

You.com Inference Router serves 12 open-weight models through 1 OpenAI-compatible API with 99.9% uptime SLA and zero retention.

Mistral Voxtral 3 adds speech-to-text, text-to-speech and a voice agent SDK, with transcription from $0.0035 per audio minute ...

Vercel open-sourced durable-harness, an Apache-2.0 package that makes Pi, OpenCode, Claude Code and Codex agent sessions ...