Particle-1 vs Kimi K3.1 Pro: 20 Real Bugs
Replays of Particle-1 vs Kimi K3.1 Pro: 20 Real Bugs: every model on every bug, side by side, with the result of each run.
Result
- Particle-1 fixed 13 of 20 bugs and scored 28 of 44 points ($0.96, 131 min 46 s).
- Kimi K3.1 Pro fixed 11 of 20 bugs and scored 23 of 44 points ($14.71, 94 min 58 s).
Bugs in this video
- Saved race answers failed
- Old interview returned after changing careers
- Home lists hid choices or overflowed
- Interview helper failed the site check
- Signed in identity was not saved
- Rivals left bends and car speeds differed
- Interview skipped ahead during teaching
- Feedback stopped before it finished
- Track edge did not slow the car
- Server startup failures were unclear
- Race car differed from chosen car
- Race car pictures did not load
- Careers reused the same answers
- Question history lacked timing
- Preview questions kept repeating
- Race track and cars looked too small
- Race difficulty choice was missing
- Profession search had no label
- Home page felt crowded
- Interview page failed to build