AiRecMark/Intelligence/Largest score gaps
D-02 · Largest score gapsarchive-recorded · snapshot 2026-09-27

Largest Score Gaps: Top 20

The 20 head-to-head dossiers with the widest recorded overall-score gaps, out of 1621 pair dossiers (of 1736 total dossiers). The widest recorded gap is 9.2 points (midjourney-vs-sora).

Key numbers

Pair dossiers
1621
data/comparisons.json (mode=pair) · N=1736
Widest gap
9.2 pts
delta · N=1621
Widest matchup
midjourney-vs-sora
delta · N=1621
Pairs at 5 pts or more
133
delta · N=1621
Median pair gap
1.6 pts
delta · N=1621

Largest gaps — top 20 of 1621 pairs

descending delta; scores from archives (score.overall)

#MatchupWinner (overall)Runner-up (overall)Gap (pts)Winner
1Midjourney vs SoraMidjourney · 85.7Sora · 76.59.2Midjourney
2Murf vs PlayHTMurf · 80.2PlayHT · 89.29PlayHT
3FLUX vs Topaz Photo AIFLUX · 88.3Topaz Photo AI · 79.98.4FLUX
4Image Creator vs FLUXImage Creator · 80.2FLUX · 88.38.1FLUX
5FLUX vs KittlFLUX · 88.3Kittl · 80.38FLUX
6FLUX vs FotorFLUX · 88.3Fotor · 80.57.8FLUX
7FLUX vs Playground AIFLUX · 88.3Playground AI · 80.57.8FLUX
8ChatGPT vs ExaChatGPT · 90.9Exa · 83.27.7ChatGPT
9F5-TTS vs PlayHTF5-TTS · 81.5PlayHT · 89.27.7PlayHT
10LALAL.AI vs PlayHTLALAL.AI · 81.6PlayHT · 89.27.6PlayHT
11Fish Audio vs PlayHTFish Audio · 81.7PlayHT · 89.27.5PlayHT
12FLUX vs PixlrFLUX · 88.3Pixlr · 80.87.5FLUX
13ChatGPT vs SciSpaceChatGPT · 90.9SciSpace · 83.57.4ChatGPT
14Coqui TTS vs PlayHTCoqui TTS · 81.8PlayHT · 89.27.4PlayHT
15FLUX vs Remove.bgFLUX · 88.3Remove.bg · 80.97.4FLUX
16Krisp vs PlayHTKrisp · 81.8PlayHT · 89.27.4PlayHT
17FLUX vs GPT ImageFLUX · 88.3GPT Image · 817.3FLUX
18Canva vs Topaz Photo AICanva · 87.1Topaz Photo AI · 79.97.2Canva
19ChatGPT vs Qwen ChatChatGPT · 90.9Qwen Chat · 83.77.2ChatGPT
20PlayHT vs SunoPlayHT · 89.2Suno · 827.2PlayHT

Ten widest pairs

midjourney vs sora9.2 ptsmurf vs playht9 ptsflux vs topaz-photo-ai8.4 ptsbing-image-creator vs flux8.1 ptsflux vs kittl8 ptsflux vs fotor7.8 ptsflux vs playground-ai7.8 ptschatgpt vs exa7.7 ptsf5-tts vs playht7.7 ptslalal-ai vs playht7.6 pts

Overall-score gap in points for the ten most one-sided recorded matchups.

Method

  • Source: data/comparisons.json, pair dossiers only (mode=pair, N=1621 of 1736 dossiers).
  • Gap = the numeric value in each dossier's delta field; participant overall scores come from the tool archives.
  • Ordering: gap descending, ties broken by dossier slug.
  • All values are archive-recorded or dossier-recorded; nothing is re-scored.
Data appendix. Source: data/comparisons.json + data/tools/*.json · snapshot 2026-09-27 · method: pairs sorted by parsed delta desc, top 20 · aggregates published with method and N. T1 benchmarks recorded: 3 / 448.