Models/Systems
D-FINE: realtime object detector - 05/05/25
- Organization: University of Science and Technology of China
- Paper: D-FINE: REDEFINE REGRESSION TASK IN DETRS AS FINE-GRAINED DISTRIBUTION REFINEMENT
- Hugging Face space
A real-time object detector much faster and accurate than YOLO with Apache 2.0 license just landed to @huggingface transformers 🔥
— merve (@mervenoyann) May 5, 2025
D-FINE is the sota real-time object detector that runs on T4 (free Colab) 🤩
Keep reading for the paper explainer, notebooks & demo 👀 pic.twitter.com/GNj2MMa8sK
Kevin-32B: Multi-Turn RL for Writing CUDA Kernels - 05/06/25
- Organization: Stanford University, Cognition AI
- Announcement
Gemini 2.5 Pro (I/O Edition) - 05/06/25
- Organization: Google
- Announcement
Very excited to share the best coding model we’ve ever built! Today we’re launching Gemini 2.5 Pro Preview 'I/O edition' with massively improved coding capabilities. Ranks no.1 on LMArena in Coding and no.1 on the WebDev Arena Leaderboard.
— Demis Hassabis (@demishassabis) May 6, 2025
It’s especially good at building… pic.twitter.com/9vRaP6RTTo
Mistral 3 medium - 05/07/25
- Organization: Mistral AI
- Announcement
- Easy-to-deploy, and state-of-the-art model at 8X lower costs.
Introducing Mistral Medium 3: our new multimodal model offering SOTA performance at 8X lower cost.
— Mistral AI (@MistralAI) May 7, 2025
- A new class of models that balances performance, cost, and deployability.
- High performance in coding and function-calling.
- Full enterprise capabilities, including hybrid or… pic.twitter.com/3j0j6X7jwj
Meta 3D AssetGen 2.0 - 05/08/25
Papers
- RM-R1: Reward Modeling as Reasoning - 05-05-25
- Towards Cross-Modality Modeling for Time Series Analytics: A Survey in the LLM Era - 05/05/25
- Absolute Zero: Reinforced Self-play Reasoning with Zero Data - 05/06/25
- ZeroSearch: Incentivize the Search Capability of LLMs without Searching - 05/07/25
- Generating Physically Stable and Buildable LEGO Designs from Text - 05/08/25
- LLMs Get Lost In Multi-Turn Conversation
Agents
- Hugging Face releases a free Operator-like agentic AI tool - 05/06/25
- From Retrieval to Reasoning: Advancing AI Agents for Knowledge Discovery and Collaboration
- Introducing web search on the Anthropic API
Products
- OpenAI releases Reinforcement Fine-tuning (RFT) for the OpenAI o4-mini allowing users to improve model performance on specific tasks
- Windsurf Wave 8: Teams & Enterprise Features
- AI-generated code could be a disaster for the software supply chain
Health
- Building foundation models for understanding human pathology using DINOv2
- Finch- AI agent that fully automates data-driven discovery in biology