Six Navers Lab / Einsia benchmarks and the work they test
Compare the tasks, scoring and evidence behind six Einsia benchmarks, with complete Radar detail pages and original references.
Model reports
Brain-to-Text '25
Evaluates one participant (T15), with about two thirds of 1,450 held-out sentences used for the private leaderboard. Speaking strategy, language model and corpus mix affect word error rate, so these scores are not directly comparable to Brain-to-Text '24.
Evidence available for this benchmark
This source record contains 6 numeric observations. The source distinguishes 6 model identities or configurations. A model may appear more than once; check the version, tools and run budget before comparing. Read results and row-level evidence →
- Reported metric
- word_error_rate
- Recorded unit
- ratio
- Metric direction
- lower is better
Start with these resources
Cited documents and dates (1)
Evaluation context and evidence
- Source
- Model reports
- Publisher
- Not recorded
- Benchmark release
- 2025-07-07
- Collected
- Not recorded
- Modality
- Not recorded
These are source-reported observations, not independent reruns by this site. Check the task version, evaluation protocol, unit and reporting date before comparing models. Results from different setups are not a universal ranking.
Reported results 6
Shown in source observation order, not ranked. Dates retain the precision provided by the source.
| Model | Value | Unit | Protocol | Reported date and basis | Source |
|---|---|---|---|---|---|
| BIT submission | 0.01773 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |
| EPFL-INL submission | 0.01663 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |
| Infera-Neuro submission | 0.01537 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |
| Stanford-NPTL causal RNN Ensemble + 5gram | 0.03892 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |
| Stanford-NPTL causal RNN TTA-Ensemble + 5gram | 0.04817 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |
| UCD-NPL causal RNN + 5gram | 0.06999 | ratio | approximately 67% of 1,450 held-out test sentences; aggregate word error rate | 2026-09-19 · document_publication | Brain-to-Text '25 private leaderboard (2026-09-19 public snapshot) |