News
EveryLLM

Reasoning models

OpenAI o1 and o3, DeepSeek-R1 and the models that think step by step before they answer. 100 models on EveryLLM so far.

Reasoning models are language models trained to work through a problem in a long chain of intermediate steps before they give a final answer. The extra thinking tokens cost time and compute each time the model is used, but they lift results on mathematics, science and coding. OpenAI's o1, previewed in September 2024, made the approach mainstream, and DeepSeek-R1 in January 2025 showed that it could be reproduced with open weights.

Where the idea came from

Researchers at Google showed in a January 2022 paper that simply prompting a large model to write out its steps, which they called chain-of-thought prompting, improved its answers on arithmetic and logic problems. Reasoning models go further. They are trained with reinforcement learning, rewarded when a long chain of thought ends in a correct, checkable answer, so the habit of planning, checking and backtracking is built into the model itself.

The first wave

OpenAI released o1-preview on 12 September 2024 and reported that it solved 83% of problems on a qualifying exam for the International Mathematics Olympiad, against 13% for GPT-4o. DeepSeek-R1 followed on 20 January 2025 under the MIT licence, with a paper describing how reasoning emerged from large-scale reinforcement learning. Anthropic's Claude 3.7 Sonnet (February 2025) and later models let users choose how long the model thinks.

Costs and limits

A reasoning model can write thousands of hidden tokens for one question, so answers are slower and dearer to produce. Labs now train models that decide for themselves how much to think: GPT-5, released in August 2025, routes each request between a fast model and a deeper reasoning model. Researchers also debate how faithfully the visible chain of thought reflects what the network actually computes.

Famous reasoning models

Further reading

fnGLM-5 馃嚚馃嚦 ChinaGLM-5 Open-weight model for engineering and agentic tasksGeneral 路 Open weights 路 2026 fnERNIE-4.5-21B-A3B-Thinking 馃嚚馃嚦 ChinaERNIE-4.5-21B-A3B-Thinking Lightweight reasoning model with 21 billion parametersReasoning 路 Open weights 路 2025 fnHunyuan-A13B 馃嚚馃嚦 ChinaHunyuan-A13B Open-source MoE language model with 80 billion total parametersGeneral 路 Open weights 路 2025 imgStep3 馃嚚馃嚦 ChinaStep3 Multimodal reasoning model with 321 billion parametersMultimodal 路 Open weights 路 2025 fnSeed-OSS-36B 馃嚚馃嚦 ChinaSeed-OSS-36B Open-source 36B parameter language model with controllable reasoning.General 路 Open weights 路 2025 imgvidIntern-S1 馃嚚馃嚦 ChinaIntern-S1 Open-source multimodal reasoning model for scienceReasoning 路 Open weights 路 2025 MiMo-7B 馃嚚馃嚦 ChinaMiMo-7B Open reasoning model from Xiaomi for maths and codeReasoning 路 Open weights 路 2025 Baichuan-M2 馃嚚馃嚦 ChinaBaichuan-M2 Open-source medical reasoning model with 32 billion parameters.Reasoning 路 Open weights 路 2025 K2-Think WorldK2-Think A 32 billion parameter open reasoning modelReasoning 路 Open weights 路 2025 fnStep-3.5-Flash 馃嚚馃嚦 ChinaStep-3.5-Flash Open-source MoE model for reasoning and agentsGeneral 路 Open weights 路 2026 EXAONE Deep 馃嚢馃嚪 South KoreaEXAONE Deep Reasoning-focused language models from LG AI ResearchReasoning 路 Open weights 路 2025 imgvidDiffusionGemma 馃嚭馃嚫 United StatesDiffusionGemma A multimodal Gemma 4 model that writes text by denoising blocks of tokens.Multimodal 路 Open weights imgvidGLM-4.1V-Thinking 馃嚚馃嚦 ChinaGLM-4.1V-Thinking Open-source vision-language model with reasoning focusMultimodal 路 Open weights 路 2025 fnGranite 4.2 馃嚭馃嚫 United StatesGranite 4.2 IBM's open reasoning language model family, released August 2026Reasoning 路 Open weights 路 2026 T-Pro 2.0 馃嚪馃嚭 RussiaT-Pro 2.0 Russian-tuned reasoning model based on Qwen 3Reasoning 路 Open weights Sarvam-105B 馃嚠馃嚦 IndiaSarvam-105B Open-weight MoE model for Indian languages and reasoningGeneral 路 Open weights imgvidfnKimi K2.5 馃嚚馃嚦 ChinaKimi K2.5 Open-source multimodal agentic model with vision and reasoning.Multimodal 路 Open weights DeepSeek-R1 馃嚚馃嚦 ChinaDeepSeek-R1 Open-weight reasoning model from ChinaReasoning 路 Open weights 路 2024

The weekly digest

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