rewirebio.iobenchmarks
Evidence and investigation

Contested results

Where one study disagrees with what other studies report for the same models, and what we are doing to resolve each case.

These are automatic leads, not findings of error. Each release, a check compares how every study ranks the models it shares with other studies. It flags a study that reverses the order the others agree on, or whose gap to a shared reference model lies well outside theirs. Different tasks, populations or metrics often explain a flag.

A flag never hides a result. Flagged results stay in every table, page and download, marked Contested. A result is withheld only when a reviewer marks it disputed.

How the check works, and how each follow-up is recorded: discrepancies between studies.

In this release

  • 192open
  • 6settled
  • 7resolved
  • 3medium severity
  • 202low severity

Release 2026-11-21-c8bea25e5abb · 205 findings in all · rules version 5

Findings

Status: All · Open · Settled · Resolved

Severity: All · Medium · Low

State: All · Not yet reviewed · Under investigation · Explained · Confirmed error · Not a real disagreement

Showing findings that involve results in Assess methods for protein stability experiments, from the study that disagrees or from the agreeing group. All use cases

3 findings match, most severe first.

ESM-1b and EVE: Brandes et al. 2023 Supplementary Table 2: full benchmark evaluations of ESM1b and all other methods reverses the order

  • Low severity
  • Explained
  • Ordering reversal
  • Same task, different protocols

Brandes et al. 2023 Supplementary Table 2: full benchmark evaluations of ESM1b and all other methods ranks ESM-1b above EVE in 6 of 6 comparisons; 2 of 2 other studies rank EVE above ESM-1b.

2 papers rank EVE above ESM-1b in 22 of their 26 comparisons and ESM-1b higher in 1; 3 are ties. Brandes et al. 2023 ranks ESM-1b above EVE in all 6 comparisons that are not ties (2 are ties).

Comparisons in which each paper scores EVE or ESM-1b higher2 other papers: EVE higher in 22, ESM-1b higher in 1, 3 ties. Brandes et al. 2023 (disagrees): EVE higher in 0, ESM-1b higher in 6, 2 ties.2 other papers22 EVE higher223 tie31 ESM-1b higherDisagrees: Brandes et al. 2023Brandes et al. 2023 Supplementary Table 2: full benchmark evaluations of ESM1b and all other methods2 tie26 ESM-1b higher6
EVE higherTieESM-1b higher
Compared
ESM-1b, EVE

3 follow-up actions, latest 21 November 2026

ESM-IF1 and ProteinMPNN: ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design and proteinbench primary benchmark evidence disagree

  • Low severity
  • Explained
  • Two-study conflict
  • different-task

ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design ranks ESM-IF1 above ProteinMPNN in 19 of 19 comparisons; proteinbench primary benchmark evidence ranks ProteinMPNN above ESM-IF1 in 10 of 10 comparisons. No other study gives a verdict, so neither is the outlier.

2 papers rank ESM-IF1 above ProteinMPNN in 19 of their 29 comparisons and ProteinMPNN higher in 10. Of those papers, proteinbench primary benchmark evidence ranks ProteinMPNN higher on its own (10 of 10).

Comparisons in which each paper scores ESM-IF1 or ProteinMPNN higher2 other papers: ESM-IF1 higher in 19, ProteinMPNN higher in 10, 0 ties.2 other papers19 ESM-IF1 higher1910 ProteinMPNN higher10
ESM-IF1 higherTieProteinMPNN higher
Compared
ESM-IF1, ProteinMPNN

2 follow-up actions, latest 24 October 2026

ESM3 and ProteinMPNN: ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design and proteinbench primary benchmark evidence disagree

  • Low severity
  • Explained
  • Two-study conflict
  • different-task

ProteinGym: Large-Scale Benchmarks for Protein Fitness Prediction and Design ranks ESM3 above ProteinMPNN in 18 of 19 comparisons; proteinbench primary benchmark evidence ranks ProteinMPNN above ESM3 in 10 of 10 comparisons. No other study gives a verdict, so neither is the outlier.

2 papers rank ESM3 above ProteinMPNN in 18 of their 29 comparisons and ProteinMPNN higher in 11. Of those papers, proteinbench primary benchmark evidence ranks ProteinMPNN higher on its own (10 of 10).

Comparisons in which each paper scores ESM3 or ProteinMPNN higher2 other papers: ESM3 higher in 18, ProteinMPNN higher in 11, 0 ties.2 other papers18 ESM3 higher1811 ProteinMPNN higher11
ESM3 higherTieProteinMPNN higher
Compared
ESM3, ProteinMPNN

2 follow-up actions, latest 24 October 2026