Final Results — FRAME Track

This page is the official final ranking of the FRAME track, computed on the test phase.

The leaderboards of the pre-evaluation and test phases on this platform sort entries by mean accuracy. They are not the final ranking. Rank by mean accuracy is shown in the last column below for comparison.

How to read the leaderboard

  • Rank comes from pairwise comparisons, not from mean accuracy. For every pair of entries and every bucket, a paired bootstrap over source videos (10,000 resamples) tests whether one entry is significantly more accurate (α = 0.05). Entry A beats entry B if A wins in more buckets than B does.
  • Copeland score = number of entries beaten − number of entries lost to. Entries with the same score share a rank.
  • Tie-break win rate is used only for ties within the top three places: the average share of video-bootstrap resamples in which an entry has the highest mean accuracy within its tie group.
  • The two official baselines (in italics) are ranked like every other entry. A team is eligible for a prize if it ranks strictly above both baselines. Teams that do not are listed as "Anonymous".
  • The seed of all bootstraps, 67, comes from random.org (integer generator, min 1, max 100, drawn 2026-09-23 10:45:31 UTC) and was fixed before this ranking was computed.

FRAME

19 teams were ranked together with the two official baselines; 17 of them rank above both baselines.

Rank Team Copeland score Opponents beaten / lost to Tie-break win rate Mean accuracy (%) Rank by mean accuracy Links
1 VLM_MAXXING +18 18 / 0 0.387 68.3 1 Team page · GitHub
2 FOKillerTeam +18 18 / 0 0.374 67.2 3 Team page · GitHub · Hugging Face
3 CBH +18 18 / 0 0.239 67.9 2 Team page · GitHub
4 Jmees +14 17 / 3 66.1 5 Team page · GitHub
5 il-surgical +11 15 / 4 66.6 4 Team page · Google Drive
6 SIMILAB-AVAHI +10 14 / 4 64.9 6 Team page · GitHub
7 UCF_IAI_Knights +6 11 / 5 64.6 7 Team page · GitHub · Hugging Face
7 Mexico-Oxford_TEAM +6 12 / 6 63.3 8 Team page · GitHub · Hugging Face
9 UTN-FunAI +3 10 / 7 62.7 9 Team page · GitHub
9 HeyDonto Labs +3 11 / 8 62.2 10 Team page · Hugging Face
11 Thunderbolts +2 9 / 7 61.8 11 Team page · GitHub · Hugging Face
12 mlo-lab -2 9 / 11 61.4 12 Team page · GitHub
13 Camma -4 8 / 12 60.9 13 Team page · GitHub · Hugging Face
14 GenMI-Star -6 6 / 12 60.2 14 Team page · Hugging Face
15 SLAI -7 6 / 13 60.0 15 Team page · GitHub
16 xAILab Bamberg -11 4 / 15 57.4 17 Team page · GitHub · Hugging Face
17 Episomic-CBBL -12 4 / 16 58.3 16 Team page · GitLab
— above: ranked higher than both baselines —
18 Official fine-tuned baseline -13 3 / 16 56.9 18
19 Anonymous -16 2 / 18 53.4 19
20 Anonymous -18 1 / 19 43.4 20
21 Official proprietary baseline -20 0 / 20 34.5 21

Accuracy per bucket (%)

Rank Team Object recognition (ID) Object recognition (OOD) Aggregation (ID) Aggregation (OOD)
1 VLM_MAXXING 72.2 56.2 74.1 70.5
2 FOKillerTeam 75.8 52.7 74.2 66.1
3 CBH 73.5 54.1 72.9 71.2
4 Jmees 74.4 48.7 71.9 69.6
5 il-surgical 70.6 58.9 68.8 68.0
6 SIMILAB-AVAHI 72.7 51.5 71.1 64.2
7 UCF_IAI_Knights 67.4 54.3 69.6 67.1
7 Mexico-Oxford_TEAM 69.5 47.3 69.1 67.2
9 UTN-FunAI 70.1 48.3 68.3 64.2
9 HeyDonto Labs 69.1 42.2 70.2 67.3
11 Thunderbolts 70.6 40.8 70.5 65.3
12 mlo-lab 66.9 44.4 68.8 65.7
13 Camma 68.6 42.0 68.0 65.1
14 GenMI-Star 68.9 40.1 69.6 62.2
15 SLAI 66.4 45.0 68.6 60.2
16 xAILab Bamberg 64.8 37.9 68.9 58.0
17 Episomic-CBBL 60.2 39.6 66.5 67.0
18 Official fine-tuned baseline 62.2 40.4 62.6 62.4
19 Anonymous 56.7 39.4 58.7 58.7
20 Anonymous 50.6 23.9 52.7 46.5
21 Official proprietary baseline 23.9 21.5 42.7 49.7

Stability

Re-running the whole ranking with 5 other seeds (1, 2, 3, 4, 5) moves these entries; every other rank is unchanged. No team's prize eligibility changes under any of them. These re-runs are reported for transparency only and do not change the official result.

Official rank Team Rank under another seed Seeds that differ
11 Thunderbolts 9 1 of 5
15 SLAI 14 1 of 5

Data and decisions

  • Withdrawn submissions and submissions without a method description are not ranked.
  • Links point to each team's page on this platform and to the repository the team gave for this track (with its test-phase submission, by email or in its method description), where it gave one.