LLM Models
Browse the world's large language models. Compare parameters, benchmarks, VRAM and more.
[!NOTE] Note: " -Paddle " models use PaddlePaddle weights, while " -PT " models use Transformer-style PyTorch weights.
AquilaMoE: Efficient Training for MoE Models with Scale-Up and Scale-Out Strategies Language Foundation Model & Software Team Beijing Academy of Artificial Intelligence (BAAI) [Paper(released soon)] [Code] [github]
AquilaMoE: Efficient Training for MoE Models with Scale-Up and Scale-Out Strategies Language Foundation Model & Software Team Beijing Academy of Artificial Intelligence (BAAI) [Paper(released soon)] [Code] [github]
Ring-lite is a lightweight, fully open-sourced MoE (Mixture of Experts) LLM designed for complex reasoning tasks. It is built upon the publicly available Ling-lite-1.5 model, which has 16.8B parameters with 2.75B activated parameters.. We use a joint training pipeline combining k…
Step 3.5 Flash (visit website) is our most capable open-source foundation model, engineered to deliver frontier reasoning and agentic capabilities with exceptional efficiency. We also open-sourced the training codebase (SteptronOss), with support for continue pretrain, SFT, RL (W…
We introduce LongCat-Flash-Prover , a flagship $560$-billion-parameter open-source Mixture-of-Experts (MoE) model that advances Native Formal Reasoning in Lean4 through agentic tool-integrated reasoning (TIR). We decompose the native formal reasoning task into three independent f…
We introduce and release LongCat-Flash-Thinking , which is a powerful and efficient large reasoning model (LRM) with 560 billion total parameters, featuring an innovative Mixture-of-Experts (MoE) architecture. The model incorporates a dynamic computation mechanism that activates…
[2026.1.28] We have provided the TileLang kernels supporting prefill (chunked-prefill as well) and decode (multi-token prediction as well). The full attention version is placed at flash mla interface.py while the streaming sparse attention version is placed at streaming sparse at…
Model Introduction We introduce LongCat-Flash, a powerful and efficient language model with 560 billion total parameters, featuring an innovative Mixture-of-Experts (MoE) architecture. The model incorporates a dynamic computation mechanism that activates 18.6B∼31.3B parameters (a…
We introduce an updated version of LongCat-Flash-Thinking, named LongCat-Flash-Thinking-2601, a powerful and efficient Large Reasoning Model (LRM) with 560 billion total parameters, built upon an innovative Mixture-of-Experts (MoE) architecture.
[!NOTE] Note: " -Paddle " models use PaddlePaddle weights, while " -PT " models use Transformer-style PyTorch weights.
[!NOTE] Note: " -Paddle " models use PaddlePaddle weights, while " -PT " models use Transformer-style PyTorch weights.