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

Mixture of experts

Mixtral, DeepSeek-V3, Kimi K2: many expert networks, a few used for each token. 112 models on EveryLLM so far.

A mixture of experts (MoE) model replaces each dense feed-forward block with many smaller expert networks and a router that sends each token to only a few of them. The model can hold a very large number of parameters while using only a fraction for any one token, so it learns more without costing proportionally more to run. Mixtral, DeepSeek-V3, Qwen3, Llama 4, gpt-oss and Kimi K2 are all built this way.

Total and active parameters

An MoE has two sizes. DeepSeek-V3 (December 2024) holds 671 billion parameters but activates 37 billion for each token. Mixtral 8x7B (December 2023) has eight experts per layer, uses two for each token, and has 46.7 billion parameters in total with 12.9 billion active. Kimi K2 (July 2025) reaches one trillion in total with 32 billion active. All the experts must still sit in memory, so MoE models need large servers even when they are fast.

A long history

The idea dates to 1991, and Google researchers led by Noam Shazeer revived it for language in 2017 with a sparsely gated layer of thousands of experts. Google's Switch Transformer reached 1.6 trillion parameters in 2021. Open MoE models took off with Mixtral, then Databricks' DBRX (March 2024, 132 billion total, 36 billion active), and became the norm for large Chinese open models. Training them is harder, because routers must spread tokens evenly across experts.

Famous mixture of experts models

Further reading

imgLlama 🇺🇸 United StatesLlama Meta AI's family of open-weight language models, from 1B to 2T parametersGeneral · Conditional licence · 2023 imgWu Dao 🇨🇳 ChinaWu Dao BAAI's trillion-parameter multimodal model family, announced 2021Multimodal · Research only · 2021 imgwavvidGemini 1.5 Pro 🇺🇸 United StatesGemini 1.5 Pro Google DeepMind's mid-size multimodal model with up to a 1 million token context windowGeneral · API only · 2024 Switch Transformer WorldSwitch Transformer A sparse mixture-of-experts language model scaling to a trillion parameters.Research · Research only · 2021 GLaM 🇺🇸 United StatesGLaM A trillion-parameter mixture-of-experts language model from Google ResearchResearch · Research only · 2021 NLLB-200 🇺🇸 United StatesNLLB-200 A research translation model covering 200 languages, built for low-resource ones.General · Conditional licence · 2022 SUTRA WorldSUTRA Multilingual model separating concepts from language processingGeneral · Research only · 2024 PanGu-Sigma 🇨🇳 ChinaPanGu-Sigma Trillion-parameter sparse language model for Chinese NLP tasksGeneral · Research only · 2023 fnNorth Mini Code 🇺🇸 United StatesNorth Mini Code A 30B sparse mixture-of-experts model for code and terminal agent workCode · Conditional licence MAI-1-preview 🇺🇸 United StatesMAI-1-preview Microsoft AI's first in-house foundation model, a mixture-of-experts system for everyday queriesGeneral · API only · 2025 imgMM1 WorldMM1 A family of multimodal language models up to 30B parameters, dense and MoE.Multimodal · Research only · 2024

The weekly digest

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