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

fnDeepSeek 🇨🇳 ChinaDeepSeek A Chinese series of large language models and a free chatbot appGeneral · API only · 2025 imgwavvidGemini 🇺🇸 United StatesGemini Google DeepMind's multimodal LLM family, from Nano to UltraGeneral · API only · 2023 imgwavvidQwen 🇨🇳 ChinaQwen A family of large and small language models developed by Alibaba Cloud.General · Open weights · 2023 imgAlice AI 🇷🇺 RussiaAlice AI Yandex's family of language, vision and image models behind the Alice assistantGeneral · API only · 2023 imgLlama 🇺🇸 United StatesLlama Meta AI's family of open-weight language models, from 1B to 2T parametersGeneral · Conditional licence · 2023 imgHuawei PanGu 🇨🇳 ChinaHuawei PanGu Huawei's multimodal large model family for industry, from 1B to trillion-scaleMultimodal · Open weights · 2023 DBRX WorldDBRX A 132-billion-parameter mixture-of-experts language model from Databricks.General · Conditional licence · 2024 imgfnMuse 🇺🇸 United StatesMuse Meta's multimodal LLM family, successor to Llama, with open and closed variantsGeneral · API only · 2026 imgWu Dao 🇨🇳 ChinaWu Dao BAAI's trillion-parameter multimodal model family, announced 2021Multimodal · Research only · 2021 fnGLM 🇨🇳 ChinaGLM Open-weight Chinese model series used for chat and codingGeneral · Open weights · 2021 imgwavvidGemma 🇺🇸 United StatesGemma Google DeepMind's open family of Gemini-derived language modelsGeneral · Open weights · 2024 imgwavvidNemotron 🇺🇸 United StatesNemotron Nvidia family of language and multimodal models for enterprise and reasoningGeneral · Conditional licence · 2023 imgwavInkling WorldInkling An open-weights mixture-of-experts model handling text, images and audio inputGeneral · Open weights · 2026 fnKimi K2 Thinking 🇨🇳 ChinaKimi K2 Thinking Open-source reasoning model with tool use and INT4 quantizationReasoning · Open weights Qwen2.5-Max 🇨🇳 ChinaQwen2.5-Max Large-scale mixture of experts model by Alibaba CloudGeneral · API only · 2025 ERNIE 5.5 Flash 🇨🇳 ChinaERNIE 5.5 Flash Efficient sparse MoE language model with 280B total parametersGeneral · Open weights MiMo-V2-Flash 🇨🇳 ChinaMiMo-V2-Flash High-speed MoE language model for reasoning and agentsGeneral · Open weights PLLuM 🇵🇱 PolandPLLuM Polish-focused large language model familyGeneral · Open weights · 2024 Snowflake Arctic 🇺🇸 United StatesSnowflake Arctic Open enterprise language model with hybrid MoE designGeneral · Open weights · 2024 Mixtral 8x7B 🇫🇷 FranceMixtral 8x7B Sparse mixture-of-experts language model with open weightsGeneral · Open weights · 2023 fnNemotron 3 Super 🇺🇸 United StatesNemotron 3 Super NVIDIA's 120B hybrid MoE language model with 12B active parameters and 1M-token contextGeneral · Conditional licence · 2026 fnParam-2-17B 🇮🇳 IndiaParam-2-17B Multilingual MoE model for Indian languages and reasoningGeneral · Open weights DeepSeekMoE 🇨🇳 ChinaDeepSeekMoE Mixture-of-experts language model architecture by DeepSeek AIGeneral · Open weights · 2024 SUTRA WorldSUTRA Multilingual model separating concepts from language processingGeneral · Research only · 2024 DeepSeek-Coder-V2 🇨🇳 ChinaDeepSeek-Coder-V2 Open-source mixture-of-experts model for code and mathCode · Open weights · 2024 fnTrinity Large 🇺🇸 United StatesTrinity Large 398B-parameter sparse MoE model family by Arcee AIGeneral · Open weights · 2026 fnDeepSeek-V3.1 🇨🇳 ChinaDeepSeek-V3.1 Hybrid thinking and non-thinking language model with 671 billion parameters.General · Open weights fnDeepSeek-V3.2 🇨🇳 ChinaDeepSeek-V3.2 Efficient reasoning and agentic AI model with sparse attentionReasoning · Open weights · 2025 fnDeepSeek-V4-Pro 🇨🇳 ChinaDeepSeek-V4-Pro 1.6T parameter MoE model with one million token contextGeneral · Open weights · 2026 DeepSeek-V4-Flash 🇨🇳 ChinaDeepSeek-V4-Flash Efficient mixture-of-experts model with one million token contextGeneral · Open weights · 2026 Qwen1.5 🇨🇳 ChinaQwen1.5 Open-weight language model series with nine sizes and strong multilingual support.General · Open weights · 2024 fnK-EXAONE 2.0 🇰🇷 South KoreaK-EXAONE 2.0 Open-weight multilingual language model by LG AI ResearchGeneral · Open weights fnQwen3-235B-A22B 🇨🇳 ChinaQwen3-235B-A22B Mixture-of-experts model with switchable reasoning modesGeneral · Open weights · 2025 A.X K1 🇰🇷 South KoreaA.X K1 Sparse MoE model with controllable reasoning depthGeneral · Open weights · 2026 fnQwen3-Next-80B-A3B 🇨🇳 ChinaQwen3-Next-80B-A3B An 80-billion parameter hybrid model with ultra-long context support.General · Open weights imgfnQwen3.5-397B-A17B 🇨🇳 ChinaQwen3.5-397B-A17B A large multimodal language model with sparse activation.General · Open weights Solar Open 100B 🇰🇷 South KoreaSolar Open 100B 102B parameter Mixture-of-Experts model by UpstageGeneral · Conditional licence · 2026 fnSolar Open 2 🇰🇷 South KoreaSolar Open 2 250B parameter open-weight model for agentic tasksGeneral · Open weights · 2026 fnKanana-2 🇰🇷 South KoreaKanana-2 Open-source multilingual model for agentic AI and reasoning.General · Open weights imgvidfnKimi K2.6 🇨🇳 ChinaKimi K2.6 Open-source multimodal agentic model by Moonshot AIGeneral · Open weights fnKimi K2 🇨🇳 ChinaKimi K2 Open mixture-of-experts model for agentic tasks and codingGeneral · Open weights · 2025 fnGLM-4.5 🇨🇳 ChinaGLM-4.5 Open-source mixture of experts model for agents and reasoningGeneral · Open weights · 2025 Kimi Linear 🇨🇳 ChinaKimi Linear Hybrid linear attention model with 48 billion parametersGeneral · Open weights · 2025 imgfnLing-1T 🇨🇳 ChinaLing-1T Trillion-parameter efficient reasoning model by InclusionAIGeneral · Open weights fnLongCat-2.0 🇨🇳 ChinaLongCat-2.0 A 1.6 trillion parameter MoE language model from Meituan.General · Open weights MiniMax-Text-01 🇨🇳 ChinaMiniMax-Text-01 A large language model with a four million token context window.General · Open weights · 2025 fnLongCat-Flash 🇨🇳 ChinaLongCat-Flash Efficient 560B parameter MoE model for agentic tasksGeneral · Open weights · 2025 wavfnSarvam-30B 🇮🇳 IndiaSarvam-30B Open-weight MoE model for Indian languages and codingGeneral · 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.