Website: https://github.com/deepseek-ai/DeepSeek-LLM

Options: Local/Offline

Price: Free (downloadable)

The Deepseek LLMs are open-source models trained on both Chinese and English sources. Their largest LLM is 67B parameters with strong coding, math, and reasoning abilities. If your hardware can’t handle 67B parameters, there is also a smaller 7B parameter version you can download.

Deepseek also recently released a 1.3B parameter multimodal LLM called Janus. This smaller LLM has image recognition and generation capabilities.

Some more info from their website:

• Superior General Capabilities: DeepSeek LLM 67B Base outperforms Llama2 70B Base in areas such as reasoning, coding, math, and Chinese comprehension.

• Proficient in Coding and Math: DeepSeek LLM 67B Chat exhibits outstanding performance in coding (HumanEval Pass@1: 73.78) and mathematics (GSM8K 0-shot: 84.1, Math 0-shot: 32.6). It also demonstrates remarkable generalization abilities, as evidenced by its exceptional score of 65 on the Hungarian National High School Exam.

• Mastery in Chinese Language: Based on our evaluation, DeepSeek LLM 67B Chat surpasses GPT-3.5 in Chinese.