OOGIRI BENCH / AI EXPERIMENT
One joke.
Which AI wrote it?
Five Japanese prompts. One main answer chosen by each AI, plus two alternatives. Reveal the joke first and the model afterward.
Compare the answers (Japanese) →
Same prompts, saved responses
The demo contains real saved responses. It does not call model APIs from your browser. Ratings reflect your own preferences, not a universal model ranking. The source is currently maintained in a private repository.