/abcompare — LLM Output Comparator
Paste two model outputs and compare them — word-level diff, similarity and length, or run a blind A/B taste-test and keep score. Everything stays in your browser.
0%
similarity
0
words A
0
words B
0
Δ words
removed from A added in B
Word-level diff (A → B)
The two outputs are shown as Left and Right in a random order (which is A and which is B is hidden). Pick the better one, then reveal — your verdict is added to the tally so you can judge many pairs without bias.
Left
Right
Tally — A: 0 · Tie: 0 · B: 0 (0 rounds)