LiteLLM Discloses Malicious 1.107.0 Artifacts on PyPI, Docker Hub and GHCR
LiteLLM's security director says malicious 1.107.0 artifacts were published to PyPI, Docker Hub and GHCR around 4:40 a.m. PDT and detected at 6:32 a.m. PDT; the…
Global AI Digest
-
-
GMI Cloud Supports NVIDIA OpenShell Reference Architecture on Vera Rubin
GMI Cloud announced support for the NVIDIA OpenShell Reference Architecture on its infrastructure built on the NVIDIA Vera Rubin platform, combining model compu…
-
Amber Notes Cuts AI Cost 40% Per Meeting With Requesty Integration
Amber Notes routed all of its meeting-summary LLM calls through a single Requesty integration, reporting 40% lower AI cost per meeting, 99.9% AI feature availab…
-
OpenAI Models Help Uncover Critical PHP Security Flaws
OpenAI's Patch the Planet initiative used its models to find critical security flaws in PHP, which powers roughly 70% of websites, and coordinated fixes with ma…
-
GPT-Rosalind Agents Surface New Neurodegeneration Leads in Mouse-Brain Data
OpenAI's GPT-Rosalind, working with PerturbAI and Scripps Research, used thousands of parallel agents to analyze mouse-brain gene-expression data, proposing mec…
-
Papers with Code PDF Reader Adds RAG-Style Source Citations via Chandra-OCR 2
Papers with Code's PDF reader now cites sources in a full RAG-style workflow, powered by the openly available Chandra-OCR 2 model, which returns bounding boxes…
-
Sentence Transformers v6.2.0 Adds Model Merging via MergeKit and Faster Loading
Sentence Transformers v6.2.0 is released with model merging support via Arcee AI's mergekit, faster model loading, automatic cross-GPU evaluator sharing, lower-…
-
Mistral AI and NVIDIA Partner to Bring AI to Industrial Engineering
Mistral AI announced a collaboration with NVIDIA targeting industrial engineering, covering faster simulation, open benchmarks and datasets for physics AI, and…
-
Requesty Becomes Native Provider Plugin in Hermes Agent
Requesty is now a native provider plugin in Hermes Agent, letting users set a fallback policy so agents switch to the next model when a provider goes down, with…
-
B.AI Launches Discounted Model Matrix Featuring DeepSeek-V4.1-Flash-1017
B.AI's updated discount matrix, effective Oct 22 at 17:00 SGT, offers DeepSeek-V4.1-Flash-1017 at 40% off, Qwen3.8-Flash at 20% off and GLM-5.3-Flash at 30% off…
