PulseAugur
EN
LIVE 07:47:49

DeepSeek ships V4 Pro, NVIDIA builds agent model, credential leak reported

DeepSeek has quietly released its V4 Pro 0813 model, available only via API. Separately, NVIDIA has developed a 30 billion parameter Mixture-of-Experts model specifically designed for agentic tasks. Additionally, a compromised AI package has led to the leak of terabytes of sensitive credentials. AI

IMPACT New model releases from DeepSeek and NVIDIA could influence agent development and performance benchmarks.

RANK_REASON The cluster includes a new model release from DeepSeek and a new model architecture from NVIDIA, fitting the research category. [lever_c_demoted from research: ic=1 ai=1.0]

Read on Mastodon — fosstodon.org →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

DeepSeek ships V4 Pro, NVIDIA builds agent model, credential leak reported

COVERAGE [1]

  1. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] ·

    📰 Latent — Grok 4.6, a stealth DeepSeek, and a credential leak • Grok 4.6 ships as a 500K-context agent model • Grok Bot turns Grok into an assignable 'teammate

    📰 Latent — Grok 4.6, a stealth DeepSeek, and a credential leak • Grok 4.6 ships as a 500K-context agent model • Grok Bot turns Grok into an assignable 'teammate' • DeepSeek quietly ships V4 Pro 0813, API-only • NVIDIA opens a 30B MoE built for the agent layer • Claude's new water…