DGA Core Plus Absolute Return ETF

RecursiveSI’s Flint 40B coding‑AI model: 262k‑token context, MoE, SWA, beats Qwen3.6 on TerminalBench 2.1, and runs on 3090/4090 GPUs.

Explore Prism ML’s new Ternary Bonsai 2 LLMs—2B, 4B, 9B models that run fast in browsers with WebGPU. Download now on Hugging Face.

Western open‑weights AI is rising, with Reflection AI’s Beam promising high‑efficiency coding and agentic performance as a Chinese‑free alternative.

Engram’s 2.6 B‑parameter LLM uses a 4.3 B Engram table and SVD‑compressed layers to deliver richer historical and philosophical completions after 100 M tokens—see how small‑scale models are reshaping AI cost and performance.

Baseten launches a new safety framework for open‑weight AI models, partnering with Hugging Face to evaluate and curb “abliteration” and its regulatory implications.

Open‑source Laya outperforms Jev: 421M‑parameter, 35 ms decision engine for real‑time gaming & web automation.

China Telecom’s new Xing4.0‑29B‑A4B MoE LLM delivers 256K‑token context, 4B active params, and 96% faster Ascend‑NPU training while outperforming rivals on key benchmarks.

Debate heats up over Hugging Face’s security breach—was it a rogue AI swarm or a human configuration mistake? Get the facts on the incident’s real cause.

Discover how “The Hugging Bay” emerged as a backup model hub amid Hugging Face’s censorship fears—and its impact on AI labs and oversight.

Explore the new CyberTiel 35B‑A3B model—4‑bit quantized, uncensored, and faster than Qwen 3.8‑27B. See how it outperforms Opus 4.6 on real coding tasks and what that means for secure software engineering.