News
ReleasesOVHcloud unveils Olala mini, its first in-house LLMChinaA 9B model beats a 1T model on FinQA, trained with Crusoe Managed RLReleasesThe internal metaphysics of an LLMReleasesMapping Language Model Internal Representations to Physical Laws and Scientific ReasoningResearchCan China keep rogue AI in check? Meet the start-up testing the top modelsFrontierMicrosoft aims to challenge OpenAI, Anthropic with cheaper, specialized AI modelsOpen modelsMicrosoft rolls out new first-party models and updates to agentic AI continuous improvementFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIFrontierGoogle is making it easier to opt out of Gemini AI model training – here’s howFrontierOpenAI Unveils New Astra AI Model After Scrapping Prior ReleaseReleasesOVHcloud unveils Olala mini, its first in-house LLMChinaA 9B model beats a 1T model on FinQA, trained with Crusoe Managed RLReleasesThe internal metaphysics of an LLMReleasesMapping Language Model Internal Representations to Physical Laws and Scientific ReasoningResearchCan China keep rogue AI in check? Meet the start-up testing the top modelsFrontierMicrosoft aims to challenge OpenAI, Anthropic with cheaper, specialized AI modelsOpen modelsMicrosoft rolls out new first-party models and updates to agentic AI continuous improvementFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIFrontierGoogle is making it easier to opt out of Gemini AI model training – here’s howFrontierOpenAI Unveils New Astra AI Model After Scrapping Prior Release
EveryLLM

Dense decoder transformers

The GPT design: one stack of transformer layers predicting the next token. 308 models on EveryLLM so far.

The dense decoder transformer is the design behind GPT and most chatbots. It is a single stack of identical transformer layers that reads the tokens so far and predicts the next one, with a mask that stops each position from looking ahead. Dense means every parameter is used for every token, unlike in a mixture of experts. GPT-3, PaLM, Llama 3, Chinchilla, Gemma, Phi and OLMo are dense decoders.

Inside each layer

Each layer has two parts. Self-attention lets every token weigh every earlier token and gather what it needs; a feed-forward network then transforms each token's vector on its own. The transformer was introduced by Google researchers in the June 2017 paper Attention Is All You Need, for translation, and OpenAI's GPT-1 (2018) kept only its decoder half. Llama 3.1 405B, released in July 2024, stacks 126 such layers.

Why it won

A decoder trains on any text without labels, because the next token is always the answer to predict, and it scales smoothly: scaling laws published by OpenAI in 2020 and DeepMind in 2022 showed the error falling predictably with more parameters, data and compute. Its weak point is cost. Attention grows with the square of the context length, and every parameter must be read for every token, which is why mixtures of experts became popular.

Famous dense decoder models

Still to be written: OLMo (language model).

Further reading

fnGLM 🇨🇳 ChinaGLM Open-weight Chinese model series used for chat and codingGeneral · Open weights · 2021 fnBlenderBot 3 WorldBlenderBot 3 A 175B open-domain chatbot with internet access and long-term memoryGeneral · Open weights · 2022 ChatGLM-6B 🇨🇳 ChinaChatGLM-6B Open bilingual dialogue model with 6.2 billion parametersAssistant · Open weights Tamil-LLaMA WorldTamil-LLaMA Open Tamil language model based on LLaMA 2Regional · Open weights · 2023 fnHermes 3 🇺🇸 United StatesHermes 3 Open-weight instruct model based on Llama-3.1 405BGeneral · Open weights · 2024 Zephyr 7B 🇺🇸 United StatesZephyr 7B A 7 billion parameter chat model fine-tuned for helpfulness.Assistant · Open weights · 2023 OLMo 2 🇺🇸 United StatesOLMo 2 Open language model family with full training transparencyGeneral · Open weights · 2024 wavMoshi 🇫🇷 FranceMoshi Real-time full-duplex spoken dialogue modelMultimodal · Open weights · 2024 imgvidfnQwen2.5-VL 🇨🇳 ChinaQwen2.5-VL Vision-language model for images, documents and long videosMultimodal · Open weights · 2025 Sherkala-Chat 🇰🇿 KazakhstanSherkala-Chat Open LLM for Kazakh language tasksRegional · Open weights · 2025 Navarasa 2.0 🇮🇳 IndiaNavarasa 2.0 Instruction-tuned model for 15 Indian languages and English.Regional · Open weights imgfnUI-TARS 🇨🇳 ChinaUI-TARS Open-source multimodal agent for GUI tasksMultimodal · Open weights · 2025 fnMistral Small 3 🇫🇷 FranceMistral Small 3 A 24B open-weight model tuned for low latency and local deploymentGeneral · Open weights · 2025 T-Pro 2.0 🇷🇺 RussiaT-Pro 2.0 Russian-tuned reasoning model based on Qwen 3Reasoning · Open weights Airavata 🇮🇳 IndiaAiravata Instruction-tuned large language model for HindiRegional · Open weights · 2024 imgwavvidQwen2.5-Omni 🇨🇳 ChinaQwen2.5-Omni End-to-end multimodal model for real-time speech and video chat.Multimodal · Open weights · 2025 fnMiniMax-M2.7 🇨🇳 ChinaMiniMax-M2.7 Open-weight model for coding and agent tasksCode · Open weights Llama 2 🇺🇸 United StatesLlama 2 Meta's open family of pretrained and chat-tuned language models, 7B to 70BGeneral · Open weights · 2023 CyberAgentLM3 🇯🇵 JapanCyberAgentLM3 A 22-billion parameter Japanese and English chat model by CyberAgent.General · Open weights

The weekly digest

Every Monday: the model of the day, new pages on EveryLLM, and the conferences and anniversaries of the week ahead.