[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

@HuggingPapers "ByteDance introduces FR3E A new RL framework that tackles unstable exploration in LLM reasoning tasks. It enables more stable training and boosts accuracy on AIME24 leading to robust and structured outputs"
@HuggingPapers on X 2025-07-11 00:20:00 UTC 3712 followers, XXX engagements

"Learn how to train your LLM web agents more efficiently and close the gap with closed-source models. Read the full paper on Hugging Face:"
@HuggingPapers on X 2025-07-12 08:07:32 UTC 3730 followers, XXX engagements

"New from MetaStone-AI: MetaStone-S1 matches OpenAI o3-mini performance with just 32B parameters This groundbreaking reflective generative model redefines test-time scaling for advanced reasoning"
@HuggingPapers on X 2025-07-14 08:09:05 UTC 3738 followers, XXX engagements

"Qwen just released the Qwen3 235B Instruct model on Hugging Face This updated FP8 MoE model brings significant improvements in reasoning coding and multi-language capabilities alongside enhanced 256K long-context understanding"
@HuggingPapers on X 2025-07-22 07:05:33 UTC 3738 followers, XXX engagements

"Google DeepMind just dropped Gemini XXX Pro a thinking model that hits SoTA on frontier coding & reasoning while juggling 3-hour videos"
@HuggingPapers on X 2025-07-19 04:12:18 UTC 3728 followers, XXX engagements

"Ant Group researchers just dropped GUI-G on Hugging Face This new framework for GUI grounding uses Gaussian reward modeling transforming sparse binary classification to dense continuous optimization. Achieves state-of-the-art results with a XXXX% boost on ScreenSpot-Pro"
@HuggingPapers on X 2025-07-22 12:10:35 UTC 3738 followers, XXX engagements

"Qualcomm AI Research just dropped CSD-VAR A novel method for content-style decomposition in visual autoregressive models. It leverages VAR's scale-wise generation for superior disentanglement and creative flexibility"
@HuggingPapers on X 2025-07-22 00:21:33 UTC 3738 followers, 8164 engagements

"ByteDance Seed team unveils PyVision This new framework enables MLLMs to dynamically generate execute and refine Python-based tools for flexible interactive visual reasoning. A big leap towards more agentic AI"
@HuggingPapers on X 2025-07-13 12:08:32 UTC 3735 followers, 5635 engagements

"Dive into Franca: it uses novel nested Matryoshka clustering and a positional disentanglement strategy for efficient high-performance vision. Code and weights are fully open-source and releasing July 24th Paper:"
@HuggingPapers on X 2025-07-21 16:08:01 UTC 3738 followers, XXX engagements

"Kwai Keye-VL Technical Report just dropped on Hugging Face Kuaishou introduces an 8-billion-parameter multimodal foundation model engineered for cutting-edge short-video understanding. Achieves state-of-the-art results on public video benchmarks and maintains robust general-purpose vision-language abilities"
@HuggingPapers on X 2025-07-03 04:13:24 UTC 3723 followers, XXX engagements

"OpenGVLab just released Mono-InternVL-1.5 This new monolithic multimodal LLM is cheaper and faster achieving state-of-the-art performance with significantly reduced training costs and first-token latency"
@HuggingPapers on X 2025-07-21 20:07:54 UTC 3738 followers, XXX engagements

"Qwen just released their updated Qwen3-235B-Instruct on Hugging Face. It boasts significant gains in instruction following logical reasoning and 256K long-context understanding"
@HuggingPapers on X 2025-07-21 17:38:37 UTC 3738 followers, 8332 engagements

"Mistral AI just dropped Voxtral A X B & XX B open-source speech-understanding model that beats Whisper large-v3 GPT-4o mini and Gemini XXX Flash on transcription and understanding while costing of comparable APIs"
@HuggingPapers on X 2025-07-20 16:06:05 UTC 3738 followers, 1571 engagements

"Google Research and collaborators unveil AgentsNet A new benchmark for multi-agent LLM coordination Tests how LLMs collaborate in networks scaling up to XXX agents"
@HuggingPapers on X 2025-07-17 00:20:20 UTC 3738 followers, 9564 engagements

"Introducing Franca: the first fully open-source vision foundation model matching and surpassing proprietary SOTA benchmarks"
@HuggingPapers on X 2025-07-21 16:07:58 UTC 3738 followers, XXX engagements

"Alibaba released Lumos-1 on Hugging Face An autoregressive video generator that keeps the plain LLM backbone adds MM-RoPE and AR-DF and trains on only XX GPUs to match the best diffusion models"
@HuggingPapers on X 2025-07-19 16:06:40 UTC 3731 followers, XXX engagements

"MMHU A massive-scale multimodal benchmark for human behavior understanding in autonomous driving. 57k human motion clips & 1.73M frames from Waymo YouTube & self-collected data with rich annotations for motion trajectory intention & safety-critical labels"
@HuggingPapers on X 2025-07-17 18:41:55 UTC 3715 followers, 8234 engagements

"ByteDance's MiroMind AI just released MiroMind-M1 This open-source RLM series built on Qwen-2.5 achieves state-of-the-art mathematical reasoning. It uses a novel Context-Aware Multi-Stage Policy Optimization"
@HuggingPapers on X 2025-07-22 08:09:28 UTC 3738 followers, XXX engagements

"MindJourney A zero-fine-tune test-time scaling framework that couples vision-language models with a controllable video-diffusion world model letting the VLM walk around in imagined 3D space before answering spatial-reasoning questions"
@HuggingPapers on X 2025-07-20 12:08:04 UTC 3738 followers, XXX engagements

"New paper just dropped on Hugging Face: a statistical diagnosis for training LLM web agents. It shows how combining SFT with on-policy RL achieves SOTA performance using only XX% of the compute required by pure SFT"
@HuggingPapers on X 2025-07-12 08:07:30 UTC 3726 followers, XXX engagements

"A 165-page survey distilling 1300+ papers into Context Engineering - the discipline that moves AI from static prompts to dynamic production-grade context orchestration"
@HuggingPapers on X 2025-07-18 07:13:41 UTC 3738 followers, 1463 engagements

"Unpack Mono-InternVL-1.5 It leverages visual attention experts EViP++ and a fused CUDA kernel. Expect state-of-the-art performance with XX% faster first-token latency & reduced training costs. Paper: Models:"
@HuggingPapers on X 2025-07-21 20:07:56 UTC 3738 followers, XXX engagements

"NVIDIA just released Long-RL It's a full-stack framework Scaling reinforcement learning to long videos up to 256k tokens on a single A100 node"
@HuggingPapers on X 2025-07-11 04:19:46 UTC 3731 followers, 8566 engagements

"One Token to Fool LLM-as-a-Judge A single : or Lets solve this step-by-step can dupe generative reward models into giving false-positive scoreshighlighting a critical flaw in RL pipelines"
@HuggingPapers on X 2025-07-19 08:06:20 UTC 3726 followers, XXX engagements