I built RIVAL to see how different AI models respond to the exact same one-shot prompt, side-by-side in a visual way. It lets you compare one-shot responses from models like GPT4.5, Claude Sonnet 3.7, Grok 3, and others across different challenge types (coding, web design, SVG generation, text analysis).
Live demo: rival.tips
GitHub: https://github.com/nuance-dev/rival
Feedback and contributions welcome
Fun! I am a bit sad to see that the source is no longer available.
This is great comparison tool! Thank you for your great work!
I understand we can only see comparison of predesigned prompts you created, right?
It would be great if we could get live replies from different models and compare them in real time.
I understand it would make you bankrupt very fast, if it was done via your API calls :).
But is there a way if I could log into my accounts for (Gemini, Claude, ChatGPT, Grok, Perplexity, DeepSeek, Qwen, KIMI, Z.AI, Le Chat) in one interface and only feed the prompt once to all the models to get all 10 replies in one go?
Just wanted to update. I did try Clawdbot recently. It can help you achieve maybe not exactly what I was proposing: 1 prompt - 10 LLM replies via subscription (not API) but close results via OAuth. If you are interested in this topic I recommend trying it out.