Gemini 2.0 Flash Lite
di Google DeepMind · Rilascio 2025-02-25
64.2
punteggio medio
$0.07/1M
Prezzo Input
$0.30/1M
Prezzo Output
1.0M tokens (~524 books)
Finestra di Contesto
multimodal
Tipo
Tested on 5 benchmarks with 64.2% average. Top scores: HELM — IFEval (82.4%), HELM — WildBench (79.0%), HELM — MMLU-Pro (72.0%).
Punteggi Benchmark
| Benchmark | Categoria | Punteggio | Bar |
|---|---|---|---|
| HELM — IFEval | language | 82.4 | |
| HELM — WildBench | reasoning | 79.0 | |
| HELM — MMLU-Pro | knowledge | 72.0 | |
| HELM — GPQA | knowledge | 50.0 | |
| HELM — Omni-MATH | math | 37.4 |