Codesota · Benchmark · table-recognition-challenge-mini-testHome/Leaderboards/Vision & Documents/Table Recognition/table-recognition-challenge-mini-test
Unknown

table-recognition-challenge-mini-test.

table-recognition-challenge-mini-test is a state-of-the-art machine learning benchmark indexed on Codesota. This page tracks published model results, top scores per metric, and the SOTA timeline for table-recognition-challenge-mini-test.

Paper ↗Leaderboard ↓
§ 01 · Leaderboard

Results by metric.

Found a wrong score or missing run?
Use row edits to send a sourced correction into moderation.
Add / edit result ↗Report issue ↗

Teds Simple Samples

Teds Simple Samples is the reported evaluation metric for table-recognition-challenge-mini-test. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Teds Simple Samplesverifiedpapervendorcommunityunverified

Muted rows were not state of the art when published — an earlier or same-year result already scored better.

RankModelTrustScoreYearLinksFix
01Re0
From paper: Unknown
verified98.35N/AN/ALooks wrong?
02Hierarchical Table Recognizer
Hierarchical Table Recognizer (Kawakatsu, Dec 2025). Current SOTA on evaluation set. arXiv:2512.21083.
paper98.12026Source ↗Looks wrong?
03MuTabNet
MuTabNet (ICDAR 2024), multi-cell decoder + mutual learning. Evaluation set. arXiv:2404.13268, Table 4 of arXiv:2512.21083.
paper98.012026Source ↗Looks wrong?
04Davar-Lab-OCR
ICDAR 2021 SLP Task B 1st-place winner (Hikvision Davar Lab). Evaluation set score from Table 4 of arXiv:2512.21083.
paper97.882026Source ↗Looks wrong?
05MTL-TabNet (LA)
Ly & Takasu 2023 local-attention variant (Nam23LA), best Ly variant. arXiv:2303.08648. Evaluation set score from Table 4 of arXiv:2512.21083.
paper97.772026Source ↗Looks wrong?
06MTL-TabNet (GA)
Ly & Takasu 2023 global-attention variant (Nam23GA). Evaluation set score from Table 4 of arXiv:2512.21083.
paper97.62026Source ↗Looks wrong?
07XM
ICDAR 2021 SLP Task B evaluation set. Score from Table 4 of arXiv:2512.21083.
paper97.62026Source ↗Looks wrong?
08MTL-TabNet (WS)
Ly et al. 2023 weakly-supervised variant (Nam23WS). Evaluation set score from Table 4 of arXiv:2512.21083.
paper97.512026Source ↗Looks wrong?
09VCGroup
From paper: Unknown
verified97.11N/AN/ALooks wrong?
10EDD
From paper: Unknown
verified95.07N/AN/ALooks wrong?
11Habitat-Web
From paper: Unknown
verified95.07N/AN/ALooks wrong?

Teds All Samples

Teds All Samples is the reported evaluation metric for table-recognition-challenge-mini-test. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Teds All Samplesverifiedpapervendorcommunityunverified

Muted rows were not state of the art when published — an earlier or same-year result already scored better.

RankModelTrustScoreYearLinksFix
01Hierarchical Table Recognizer
Hierarchical Table Recognizer (Kawakatsu, Dec 2025). Current SOTA on evaluation set. arXiv:2512.21083.
paper96.622026Source ↗Looks wrong?
02MuTabNet
MuTabNet (ICDAR 2024), multi-cell decoder + mutual learning. Evaluation set. arXiv:2404.13268, Table 4 of arXiv:2512.21083.
paper96.532026Source ↗Looks wrong?
03Davar-Lab-OCR
ICDAR 2021 SLP Task B 1st-place winner (Hikvision Davar Lab). Evaluation set score from Table 4 of arXiv:2512.21083.
paper96.362026Source ↗Looks wrong?
04XM
ICDAR 2021 SLP Task B evaluation set. Score from Table 4 of arXiv:2512.21083.
paper96.272026Source ↗Looks wrong?
05MTL-TabNet (LA)
Ly & Takasu 2023 local-attention variant (Nam23LA), best Ly variant. arXiv:2303.08648. Evaluation set score from Table 4 of arXiv:2512.21083.
paper96.212026Source ↗Looks wrong?
06MTL-TabNet (GA)
Ly & Takasu 2023 global-attention variant (Nam23GA). Evaluation set score from Table 4 of arXiv:2512.21083.
paper96.172026Source ↗Looks wrong?
07MTL-TabNet (WS)
Ly et al. 2023 weakly-supervised variant (Nam23WS). Evaluation set score from Table 4 of arXiv:2512.21083.
paper95.972026Source ↗Looks wrong?
08Re0
From paper: Unknown
verified95.66N/AN/ALooks wrong?
09VCGroup
From paper: Unknown
verified95.04N/AN/ALooks wrong?
10Habitat-Web
From paper: Unknown
verified89.97N/AN/ALooks wrong?
11EDD
From paper: Unknown
verified89.97N/AN/ALooks wrong?

Teds Complex Samples

Teds Complex Samples is the reported evaluation metric for table-recognition-challenge-mini-test. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.

Higher is better

Trust tiers for Teds Complex Samplesverifiedpapervendorcommunityunverified

Muted rows were not state of the art when published — an earlier or same-year result already scored better.

RankModelTrustScoreYearLinksFix
01Hierarchical Table Recognizer
Hierarchical Table Recognizer (Kawakatsu, Dec 2025). Current SOTA on evaluation set. arXiv:2512.21083.
paper95.082026Source ↗Looks wrong?
02MuTabNet
MuTabNet (ICDAR 2024), multi-cell decoder + mutual learning. Evaluation set. arXiv:2404.13268, Table 4 of arXiv:2512.21083.
paper94.982026Source ↗Looks wrong?
03XM
ICDAR 2021 SLP Task B evaluation set. Score from Table 4 of arXiv:2512.21083.
paper94.892026Source ↗Looks wrong?
04Davar-Lab-OCR
ICDAR 2021 SLP Task B 1st-place winner (Hikvision Davar Lab). Evaluation set score from Table 4 of arXiv:2512.21083.
paper94.782026Source ↗Looks wrong?
05MTL-TabNet (GA)
Ly & Takasu 2023 global-attention variant (Nam23GA). Evaluation set score from Table 4 of arXiv:2512.21083.
paper94.682026Source ↗Looks wrong?
06MTL-TabNet (LA)
Ly & Takasu 2023 local-attention variant (Nam23LA), best Ly variant. arXiv:2303.08648. Evaluation set score from Table 4 of arXiv:2512.21083.
paper94.582026Source ↗Looks wrong?
07MTL-TabNet (WS)
Ly et al. 2023 weakly-supervised variant (Nam23WS). Evaluation set score from Table 4 of arXiv:2512.21083.
paper94.372026Source ↗Looks wrong?
08VCGroup
From paper: Unknown
verified92.96N/AN/ALooks wrong?
09Re0
From paper: Unknown
verified92.96N/AN/ALooks wrong?
10EDD
From paper: Unknown
verified84.86N/AN/ALooks wrong?
11Habitat-Web
From paper: Unknown
verified84.86N/AN/ALooks wrong?
§ 04 · Submit a result

Add to the leaderboard.

← Back to Table Recognition
table-recognition-challenge-mini-test Leaderboard | CodeSOTA | CodeSOTA