AI’s TwIL-8B is an 8.8B open-weights reasoning model scoring 98.2 on MATH-500 and running locally in 4.98 GiB.

Ricursive’s Ripple-65B Preview is a 65B dense open-weight model scoring 80.5 on Terminal-Bench 2.1, released under Apache 2.0 ...

Tencent WeVisOne-VL is an open 2B/4B vision language model scoring 71.3 F1 on Dense200 without a separate SAM.

NVIDIA shows 3 ways to speed up MoE inference on DGX Spark, lifting Ling-3.0-flash GSM8K decode throughput 42.5%.

Size: Soniox STT v6 is a proprietary hosted model, no parameter count published. Runs on: Soniox cloud API only (WebSocket ...

There is no tool for browsing raw schemas. NVIDIA chose that on purpose, so agents reach data through the semantic layer. The ...

Anthropic FlashProve generates zero-knowledge proofs for 400B Llama 4 Maverick inference at 110 tokens/second on 1 H100 GPU.

OpenAI Pando, a Git-compatible monorepo backend, cut median push latency nearly 80% and serves nearly 1 PB daily.

Deep Principle released Hydrogen-1.0, a 753B open-weight scientific model post-trained from GLM-5.3, scoring 38.3% on FrontierScience-Research ...

Vercel Labs agent-desktop gives AI agents computer use across macOS, Windows, Linux with 60 accessibility-tree commands, not ...

Microsoft’s TokenEconomizer plugin cuts coding agent token costs by 18.4% to 30.9% on SWE-Bench without changing the agent.

It is the second 0.6B release in Edge0’s Audio8 TTS family. It builds on Audio8-TTS-Preview-0.6b and keeps the same ...