Last 90 days · 2299 total
TW3Cast is a time-series forecasting system leveraging lightly fine-tuned foundation models, achieving third place out of 130 entries on the GIFT-Eva…
OriginalA paper titled "PAWS: Policy-driven Agentic World Simulation" introduces a framework for policy-driven agentic world simulation to study the propagat…
OriginalA technical report introduces the Pistis model family, comprising 27B and 9B parameter multimodal large language models built upon Qwen3.6 and Qwen3.…
OriginalAnthropic has released Claude Opus 5.5 (claude-opus-5-5), which is now the default Opus model, featuring a 1M context window and updated pricing.
OriginalKimi_Moonshot's Kimi K3, an open-weight model, is now generally available and rolling out in GitHub Copilot, offering frontier-level agentic coding a…
OriginalMistral AI announced an expanded global strategic partnership with Microsoft to provide enterprises and regulated industries with controllable fronti…
Originalllama.cpp released version b11172, featuring significant Metal optimizations for sparse FA and shared memory management.
OriginalvLLM v0.30.0 has been released, featuring 762 commits and adding support for new models like DeepSeek-V4.1-Flash, which stores the entire KV in MXFP8.
OriginalOpenAI has introduced a new Data agent in ChatGPT Work, enabling users to transform company data into answers, interactive dashboards, and actionable…
OriginalUnsloth AI announced that GLM-5.3-Flash can now run locally with 3-bit quantization on 128GB RAM via Unsloth GGUF, rivaling Claude Opus 4.8 on agenti…
Original