Which benchmark
should you trust?
Tasks answer what problem you are solving. Benchmarks answer whether the evidence is still useful. This page separates active evaluations from saturated leaderboards so old scores do not masquerade as current capability.
—
Benchmarks
—
With Results
—
Result Rows
—
Verified Rows
TTS speed vs quality vs cost.
Word Error Rate (WER) evaluates quality by comparing voice outputs against ground truth transcripts. Lower scores indicate fewer deleted, inserted, or substituted words.
Still discriminates frontier systems. Use for current model comparisons.
Useful but ceiling effects or contamination risks are visible. Read successor context.
Good historical anchor, weak frontier signal. Prefer the successor benchmark.
Replaced by a cleaner, harder, or more representative evaluation artifact.
Tracked leaderboard without curated lineage status yet. Treat as coverage backlog.
A leaderboard is only useful if you know whether the benchmark is current, saturated, superseded, or still carrying the field.
Where the result rows are.
Unmapped
41
Benchmarks
310
Results
81
Verified
Vision & Documents
25
Benchmarks
258
Results
95
Verified
Code & Software Engineering
8
Benchmarks
101
Results
43
Verified
Language & Knowledge
11
Benchmarks
41
Results
0
Verified