Compare DeepSeek, GPT, Gemma, Llama, Mistral and other AI LLM models at chat.files.fm
https://chat.files.fm - here you can try the 8-billion-parameter-small Deepseek R1 8B model (one that everyone can run) and test what it can do.
For comparison, there are some other popular models available with similar number of neuron parameters - Googles Gemma2:9B, llama3:8B, Mistral 7B, Alibaba's qwen2.5:7b, Microsoft phi3:mini.
p.s. The number of parameters partly determines how "smart" the model is, or knows how to generate. For models used by OpenAI and other companies, the parameters are in the hundreds of billions or thousands of billions (GPT4 is 1500B or 1.5 trillion)
Files.fm — EU-based cloud storage