Latest stories
Agentic Context Management: A New Framework for Agent Memory and Cost
Hugging Face researchers propose Agentic Context Management (ACM), a lifecycle-based approach to agent memory that treats context...
Hugging Face Study Reveals Scaling Laws for Native Multimodal Pre-Training
A new paper from Hugging Face investigates the scaling properties of native multimodal pre-training, showing that compute-optimal...
Molt: A Scalable PyTorch-Native Framework for Agentic RL
Molt is a new open-source, PyTorch-native training framework for agentic reinforcement learning that simplifies algorithm iteratio...
Skill Self-Play: Co-Evolving Skills Push LLM Capabilities to New Heights
A new framework called Skill Self-Play (Skill-SP) enables LLMs to co-evolve skills through a proposer, solver, and dynamic skill c...
DeepSeek V4: 4 Chat Modes, Chaining, 1M Context – Pro Tips
DeepSeek's latest guide reveals how to master V4's four chat modes, chain workflows, combine file uploads with web search, and lev...
DeepSeek V4 Goes GA: Legacy API Aliases Retire July 24, Peak/Off-Peak Pricing Arrives
DeepSeek V4 reaches General Availability on July 24, 2026, retiring legacy API aliases and introducing the industry's first struct...
DeepSeek Raises $7.4B in Unprecedented Founder-Friendly Deal
DeepSeek secures $7.4B at a $50B+ valuation with a unique structure giving investors zero voting rights and a 5-year lock-up, sign...
DeepSeek V4 Guide: 4 Chat Modes, Chaining, and 1M Context Window
DeepSeek releases a comprehensive guide on maximizing V4's capabilities, including four chat modes, workflow chaining, file upload...
DeepSeek's DSpark Boosts LLM Inference Speed by Up to 85% with Speculative Decoding
DeepSeek introduces DSpark, a speculative decoding system that accelerates LLM inference by 57–85% on V4, Qwen, and Gemma models w...
DeepSeek Unveils DSpark: Speculative Decoding Boosts LLM Inference by 60–85%
DeepSeek introduces DSpark, a speculative decoding framework that accelerates LLM inference by 60–85% while preserving byte-identi...
DeepSeek Builds Custom AI Chip to Challenge Nvidia and Huawei
DeepSeek is developing its own AI inference chip to reduce dependence on Nvidia and Huawei, potentially reshaping the AI hardware...
DeepSeek V4 Reaches General Availability, Retires Legacy Aliases and Introduces Surge Pricing
DeepSeek V4 goes GA on July 24, 2026, retiring legacy API aliases and introducing peak/off-peak surge pricing based on UTC+8 time...