# Long-context Reasoning

Qwen3 is the latest version in the Tongyi Qianwen series of large language models, offering a complete suite of dense and mixture-of-experts (MoE) models, achieving breakthrough progress in reasoning, instruction following, agent capabilities, and multilingual support.

Phi-4 Reasoning Enhanced is a 14-billion parameter open-source reasoning model developed by Microsoft Research, optimized through supervised fine-tuning and reinforcement learning, focusing on advanced reasoning capabilities in mathematics, science, and programming fields.

Qwen3-30B-A3B-Base is the latest 30.5B parameter-scale Mixture of Experts (MoE) large language model in the Qwen series, supporting 119 languages and 32k context length.

IBM-Granite's 2-billion parameter instruction model supporting multilingual and long-context tasks with structured reasoning capabilities.

The Llama 4 series is a multimodal AI model developed by Meta, supporting text and image interactions, utilizing a Mixture of Experts (MoE) architecture, and delivering industry-leading performance in text and image comprehension.

Raptor-X5-UIGEN is a large language model based on the Qwen 2.5 14B multimodal architecture, specializing in UI design, minimalist coding, and content-intensive development, with enhanced reasoning capabilities and structured response generation.

Mixtral-8x7B is a pre-trained generative sparse mixture of experts model that outperforms Llama 2 70B on most benchmarks.

An encoder-decoder model based on the T5 architecture, optimized through Flan prompt tuning, supporting multilingual task processing