falcon-11B — benchmark results

TII's 11B Falcon 2 generation model (May 2024), trained on 5.5T tokens with multilingual support, released under the permissive TII Falcon License 2.0. Provider: TII. Released 2024-05-14. Access: Open.

Unified ELO 1420 ± 18, rank #1127 of 1776 rated models, from 46 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
EuroEval Portuguese NLU - SST-2 PT83.44Sentiment classification Score (%)84.3
HellaSwag82.91Accuracy (%)76.3
EuroEval Dutch NLU - DBRD90.49Sentiment classification Score (%)75.8
EuroEval Italian NLU - Sentipolc1659.88Sentiment classification Score (%)73.7
WinoGrande78.3Accuracy (%)70
Open PL LLM - RAG65.34Average RAG Score (%)66.4
EuroEval Spanish NLU - ScaLA ES25.29Linguistic acceptability Score (%)65.2
EuroEval Spanish NLU - Sentiment Headlines ES43.53Sentiment classification Score (%)63.6
EuroEval Portuguese NLU - ScaLA PT16.64Linguistic acceptability Score (%)59.3
Open PL LLM - Generative50.78Average Generative Score (%)54.4
GSM8K53.83Accuracy (%)54.3
EuroEval Portuguese NLU50.52NLU Average Score (%)53.7

Interactive version: theaggregate.ai/model?slug=falcon-11b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.