Ant Group's Ling-3.1-flash Squeezes a 560B Model Into Four GPUs
Ant Group's InclusionAI opens weights for a 560B hybrid-linear MoE model with 25B active parameters, 1M context, and MIT licensing.
// Intelligence
Llms
A focused feed of models, agents, research and open-source releases for people building with AI.
Updated
Reflection AI Drops Beam, a 501B Open-Weight Agent Model for Coding
Reflection AI opens the weights of Beam, a 501B Mixture-of-Experts model, alongside a 40+ partner ecosystem and a full agent toolchain.
Tencent's Hy Memory 2.0 Beats Claude Opus 5 at Remembering What Agents Need
Tencent Hunyuan open-sources a 30B memory writer model, a dual-system memory SDK, and MemHub to unify memory across AI tools.
Cheng Lou's Handwired Compiles JavaScript Directly Into Neural Networks Without Training
A compiler converts if-else and loops into frozen matrix weights, producing exact, differentiable neural nets that run JavaScript programs in a browser.
Meta Drops Muse Spark Open, an 816B Open-Weight Model Rivaling Closed AI Giants
Meta's new 816B parameter mixture-of-experts model targets long-horizon coding and agentic work, with open weights that fit on two Mac Studios.
Liquid AI's d1-ultra Skips Text Generation and Returns Raw Probabilities
Liquid AI's new multimodal decision model tops the Artificial Analysis Decision Index, returns calibrated probabilities with zero generated tokens, and costs $0.042 per million input tokens.
QuantTrio Shrinks Alibaba's Qwen3.6 From 70 GB to 24 GB for Workstation Deployment
QuantTrio shipped a 4-bit AWQ quantization of Qwen3.6-35B-A3B that compresses the MoE down to roughly 24GiB for vLLM deployment.
NVIDIA's Nemotron 3.5 Super VL Runs 5x Faster on Agentic Tasks
NVIDIA's new open 120B mixture-of-experts model fuses text, image, document, and video reasoning into one agent-ready backbone with 12B active parameters.
OpenBMB's MiniCPM-V 4.7 Runs a 35B Vision Model on Your Phone
OpenBMB's MiniCPM-V 4.7 packs 35B parameters into a mixture-of-experts vision model that runs on an iPhone at 13+ tokens per second.
Nagoya University's Ruri v3 Beats OpenAI Embeddings on Japanese Retrieval
A 315M-parameter Japanese embedding model built on ModernBERT-Ja tops JMTEB benchmarks while handling 8K context, outperforming OpenAI and billion-parameter rivals.
MBZUAI Ships Jais 3, Six Open Arabic Models Beating Qwen and Claude
IFM's new Arabic-first model family spans 0.9B to 375B parameters, with a flagship MoE that matches Claude Sonnet 5 on Arabic benchmarks and acts as an agent.
Baidu's ProxyAttn Cuts Long-Context AI Processing 10x Without Retraining
A training-free sparse attention method uses a handful of representative attention heads as proxies, delivering 10.3x attention speedup on long contexts.
TypeSafe AI's Jev 1.14 Brings Vision and 256K Context to Typed Decisions
TypeSafe AI ships Jev 1.14 with image inputs, a bounding box primitive, and 4x longer context, all behind opt-in danger zone headers.
Google's Gemini 4 Argon Rewrites Chromium and Frees 300 TiB With 1M-Token Outputs
Google DeepMind's new frontier model ships with a 1M token output ceiling, agentic coding wins, and verified lab discoveries in fusion, quantum, and biology.
Google DeepMind Shrinks Gemma 4 E4B to 5GB Without Losing Reasoning Power
Google released quantization-aware trained Q4_0 GGUF checkpoints of Gemma 4 E4B, delivering near-bfloat16 quality in a 5.15 GB on-device package.
Japan's NII Releases LLM-jp-4 to Beat GPT-4o on Japanese Benchmarks
Japan's National Institute of Informatics releases an 8B dense and 32B MoE model trained from scratch on 12T tokens, with the full pipeline open to inspection.
Cursor Ships Gemini 4 Argon With a 1M-Token Output Limit
Google's frontier Gemini 4 Argon model lands in Cursor at $2/$10 per million tokens, matching Claude Opus 5.5 on CursorBench coding evaluations.
GitHub Rolls Out Gemini 4 Argon to Tackle Long Agent Coding Runs
Google's newest frontier model lands in Copilot with gains across every coding benchmark and especially strong results in the terminal agent.
Liquid AI's LFM2.5-24B-A2B Runs 184 Tokens per Second on a Laptop
Liquid AI's new 24B MoE activates just 2B parameters per token, hits 184 tok/s on a MacBook, and targets fully local agents.
Artificial Analysis' Decision Index Shows Jev Beats Claude Haiku 5.5 at 11x Speed
A new leaderboard pits purpose-built decision models against LLMs on classification, tool calling, knowledge, and reasoning, with cost and latency side by side.
Cognition Adds Gemini 4 Argon to Devin at 75% off
Cognition added Google's new frontier model to Devin Desktop and CLI, where it tops Gemini scores on FrontierCode at 60.5% with a 75% launch discount.
Google DeepMind's Gemini 4 Argon Tops Vals Index With 1M-Token Window
Google's new frontier model tops the Vals Index at 68.9%, ships with a 1M token window, and launches at $2/$10 per million tokens.
AutoTrust AI's JEV-35B Turns Open-Source Decisions Into 204ms API Calls
AutoTrust AI's open-weights JEV-35B turns a Qwen3.5 MoE into a System 1 decision model that picks among 256 options in 204 ms.
Ai2's TutorMoments Benchmark Catches AI Models Helping Students Too Soon
Ai2's TutorMoments benchmark replays real math sessions to test whether LLMs scaffold learning or spoil it by over-helping students.
No stories match these filters.