Gemini 2.5 Flash Lite (Sep '25): benchmark results
Provider: Google. Released 2025-09-25. Access: API.
Aggregate Rating 1640 ± 25, rank #691 of 10916 rated models, from 43 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA LiveCodeBench | 68.78 | Pass@1 (%) | 77.3 |
| AA MMLU-Pro | 80.8 | Accuracy (%) | 71.3 |
| AA Global-MMLU-Lite - Yoruba | 65.42 | Accuracy (%) | 70 |
| AA Global-MMLU-Lite - Burmese | 80.25 | Accuracy (%) | 68.2 |
| AA Global-MMLU-Lite - Korean | 86 | Accuracy (%) | 64.5 |
| AA Global-MMLU-Lite - Swahili | 79.08 | Accuracy (%) | 63.6 |
| AA AIME 2025 | 68.67 | Accuracy (%) | 63.3 |
| AA Omniscience - Software Engineering (SWE) - Kotlin | 20 | Accuracy (%) | 63 |
| AA IFBench | 52.59 | Accuracy (%) | 62.9 |
| AA Omniscience - Software Engineering (SWE) - Swift | 36 | Accuracy (%) | 60.6 |
| AA Global-MMLU-Lite - Bengali | 83.33 | Accuracy (%) | 60 |
| AA Global-MMLU-Lite | 83.79 | Accuracy (%) | 59.1 |
Interactive version: theaggregate.ai/model?slug=gemini-2-5-flash-lite-sep-25 · How It Works · Data refreshed daily, snapshot 2026-11-08.