ALIA-40B: benchmark results

Provider: BSC. Released 2025-06-01. Access: Open.

Aggregate Rating 1447 ± 26, rank #3888 of 10916 rated models, from 239 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Ukrainian NLU - Cross Domain UK Reviews63.15Matthews correlation coefficient (%)95
EuroEval Hungarian NLU - MultiWikiQA HU67.39Character F1 (%)93.1
EuroEval Estonian NLU - Grammar ET45.27Matthews correlation coefficient (%)92.4
EuroEval Polish NLU - PoQuAD59.39Character F1 (%)91
EuroEval Czech NLU - CSFD Sentiment60.92Matthews correlation coefficient (%)89.5
EuroEval Lithuanian NLU - MultiWikiQA LT69.28Character F1 (%)89.5
EuroEval Romanian NLU - MultiWikiQA RO73.24Character F1 (%)89.3
EuroEval Latvian NLU - MultiWikiQA LV66.9Character F1 (%)89.2
EuroEval Finnish NLU - Tydiqa FI72.03Character F1 (%)88.5
EuroEval Estonian NLU - MultiWikiQA ET69.89Character F1 (%)86.7
EuroEval Norwegian NLU - Norquad72.27Character F1 (%)86.4
EuroEval German NLU - Germanquad64.16Character F1 (%)86.2

Interactive version: theaggregate.ai/model?slug=alia-40b · How It Works · Data refreshed daily, snapshot 2026-11-08.