News
ResearchOpenAI pens AI safety agreement with four firmsReleasesOVHcloud 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 improvementFrontierOpenAI鈥檚 New Hacking Model Found Dozens Of Holes In Apple TechFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAIResearchOpenAI pens AI safety agreement with four firmsReleasesOVHcloud 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 improvementFrontierOpenAI鈥檚 New Hacking Model Found Dozens Of Holes In Apple TechFrontierMicrosoft unveils AI models for coding, cybersecurity as it takes on Anthropic, OpenAI
EveryLLM

Text

Reads and writes text. 471 models on EveryLLM so far.

Text is the base modality of every language model: it reads a prompt and writes a reply, token by token. A tokenizer first splits text into tokens, which are whole words, pieces of words or single characters; in English a token is roughly three-quarters of a word. Everything else a model does, from code to reasoning to calling tools, is ultimately expressed as text that it reads or writes.

Tokens and context

Most models use byte-pair encoding, which builds a vocabulary of frequent character sequences. Vocabularies have grown from about 50,000 tokens in GPT-2 to around 200,000 in GPT-4o and 256,000 in Gemma. The context window is how many tokens the model can consider at once: GPT-3 handled 2,048 in 2020, while Gemini 1.5 Pro offered one million in 2024. Languages in non-Latin scripts often need more tokens for the same meaning.

What text-only still means

Many models remain text-only, especially smaller ones and older families such as GPT-3, Llama 2 and BERT, and text still carries most business uses: drafting, summarising, pulling fields out of documents, translation and search. Text-only models are cheaper to train and serve, and easier to evaluate, since their inputs and outputs can be checked and compared directly.

Famous text models

Further reading

Zephyr 7B 馃嚭馃嚫 United StatesZephyr 7B A 7 billion parameter chat model fine-tuned for helpfulness.Assistant 路 Open weights 路 2023 OLMo 2 馃嚭馃嚫 United StatesOLMo 2 Open language model family with full training transparencyGeneral 路 Open weights 路 2024 Poro 2 馃嚝馃嚠 FinlandPoro 2 Finnish and English instruction-following model based on Llama 3.1General 路 Conditional licence wavMoshi 馃嚝馃嚪 FranceMoshi Real-time full-duplex spoken dialogue modelMultimodal 路 Open weights 路 2024 imgvidfnQwen2.5-VL 馃嚚馃嚦 ChinaQwen2.5-VL Vision-language model for images, documents and long videosMultimodal 路 Open weights 路 2025 fnQwen3-Coder 馃嚚馃嚦 ChinaQwen3-Coder Agentic coding model with 480 billion parametersCode 路 Open weights 路 2025 EXAONE 3.5 馃嚢馃嚪 South KoreaEXAONE 3.5 LG AI Research open language model series for real-world tasksGeneral 路 Conditional licence 路 2024 Sherkala-Chat 馃嚢馃嚳 KazakhstanSherkala-Chat Open LLM for Kazakh language tasksRegional 路 Open weights 路 2025 Navarasa 2.0 馃嚠馃嚦 IndiaNavarasa 2.0 Instruction-tuned model for 15 Indian languages and English.Regional 路 Open weights fnLongCat-Flash 馃嚚馃嚦 ChinaLongCat-Flash Efficient 560B parameter MoE model for agentic tasksGeneral 路 Open weights 路 2025 fnMiniMax-M2.5 馃嚚馃嚦 ChinaMiniMax-M2.5 Cost-efficient coding and agentic model from MiniMax AICode 路 API only 路 2026 imgfnUI-TARS 馃嚚馃嚦 ChinaUI-TARS Open-source multimodal agent for GUI tasksMultimodal 路 Open weights 路 2025 fnMistral Small 3 馃嚝馃嚪 FranceMistral Small 3 A 24B open-weight model tuned for low latency and local deploymentGeneral 路 Open weights 路 2025 Sahabat-AI 馃嚠馃嚛 Indonesia and 1 moreSahabat-AI Indonesian language model family by GoTo and IndosatRegional 路 Conditional licence imgLLaVA 馃嚭馃嚫 United StatesLLaVA Open-source multimodal chatbot connecting vision encoders to language models.Multimodal 路 Fully open 路 2023 fnMiniMax-M2.7 馃嚚馃嚦 ChinaMiniMax-M2.7 Open-weight model for coding and agent tasksCode 路 Open weights CyberAgentLM3 馃嚡馃嚨 JapanCyberAgentLM3 A 22-billion parameter Japanese and English chat model by CyberAgent.General 路 Open weights imgwavvidQwen2.5-Omni 馃嚚馃嚦 ChinaQwen2.5-Omni End-to-end multimodal model for real-time speech and video chat.Multimodal 路 Open weights 路 2025 OpenAssistant 馃嚛馃嚜 GermanyOpenAssistant An open source chatbot built by volunteers and LAION.Assistant 路 Fully open 路 2023 fnDevin 馃嚭馃嚫 United StatesDevin Autonomous software engineering agent by Cognition LabsCode 路 API only imgwavChatGPT 馃嚭馃嚫 United StatesChatGPT Generative AI chatbot by OpenAIAssistant 路 API only 路 2022

The weekly digest

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