DeepSeek-R1
Description
DeepSeek's first-generation reasoning models, DeepSeek-R1-Zero and DeepSeek-R1. DeepSeek-R1-Zero, a model trained via large-scale reinforcement learning (RL) without supervised fine-tuning (SFT) as a preliminary step, demonstrated remarkable performance on reasoning.
Related Tools
DeepSeek-V3
A strong Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.
Llama 3
Llama3 is a large language model developed by Meta AI. It is the successor to Meta's Llama2 language model. <br>Online test address:<br>[huggingface.co/Meta-Llama-3-70B-Instruct](https://huggingface.co/chat/models/meta-llama/Meta-Llama-3-70B-Instruct)
grok-1
A large language model open sourced by xAI
Phi-3
Phi-3, a family of open AI models developed by Microsoft. Phi-3 models are the most capable and cost-effective small language models (SLMs) available, outperforming models of the same size and next size up across a variety of language, reasoning, coding, and math benchmarks.
Gemma 4
Gemma 4 is Google's latest open source large language model series, built on the Gemini architecture, offering improved performance, longer context windows, and better multilingual support.
Qwen3
Qwen3 is the large language model series developed by the Qwen team at Alibaba Cloud. Qwen3-2507 includes Instruct and Thinking variants with 235B-A22B, 30B-A3B, and 4B sizes, 256K long context, and 1M-token input support in some scenarios.