Gemini 2.5 Flash Lite (Non-reasoning): benchmark results
Gemini 2.5 Flash Lite evaluated with reasoning disabled. Provider: Google. Released 2025-06-17. Access: API.
Aggregate Rating 1553 ± 10, rank #869 of 2170 rated models, from 416 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| EuroEval Lithuanian NLU - Atsiliepimai | 50.18 | Matthews correlation coefficient (%) | 98 |
| EuroEval Bosnian NLU - MMS BS | 52.69 | Matthews correlation coefficient (%) | 97.2 |
| EuroEval Italian NLU - SQuAD IT | 74.89 | Character F1 (%) | 96 |
| EuroEval Hungarian Summarization - Hunsum | 33.65 | chrF3++ (%) | 95.9 |
| EuroEval Swedish NLU - ScaLA SV | 69.03 | Matthews correlation coefficient (%) | 94.1 |
| EuroEval Faroese NLU - FoQA | 76.6 | Character F1 (%) | 94 |
| EuroEval Albanian Summarization - LR SUM SQ | 34.05 | chrF3++ (%) | 93.5 |
| EuroEval Ukrainian NLU - ScaLA UK | 42.26 | Matthews correlation coefficient (%) | 93.5 |
| EuroEval Dutch NLU - DBRD | 92.18 | Matthews correlation coefficient (%) | 93 |
| EuroEval French Summarization - Orange SUM | 38.62 | chrF3++ (%) | 93 |
| HELM Arabic - Aratrust - Illegal | 98.11 | Exact Match (%) | 92.5 |
| HELM Arabic - Aratrust - Unfairness | 96.36 | Exact Match (%) | 92.5 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-flash-lite-non-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-05.