Gemini 2.5 Flash-Lite
Multimodal
Zero-eval
#1MRCR v2
#1Arc
#3FACTS Grounding
by Google
+
+
+
+
About
Gemini 2.5 Flash-Lite is a multimodal language model developed by Google. The model shows competitive results across 13 benchmarks. It excels particularly in FACTS Grounding (84.1%), Global-MMLU-Lite (81.1%), MMMU (72.9%). With a 1.1M token context window, it can handle extensive documents and complex multi-turn conversations. The model is available through 1 API provider. As a multimodal model, it can process and understand text, images, and other input formats seamlessly. Released in 2025, it represents Google's latest advancement in AI technology.
+
+
+
+
Pricing Range
Input (per 1M)$0.10 -$0.10
Output (per 1M)$0.40 -$0.40
Providers1
+
+
+
+
Timeline
AnnouncedJun 17, 2025
ReleasedJun 17, 2025
Knowledge CutoffJan 1, 2025
+
+
+
+
Specifications
Capabilities
Multimodal
+
+
+
+
License & Family
License
Creative Commons Attribution 4.0 License
Performance Overview
Performance metrics and category breakdown
Overall Performance
13 benchmarks
Average Score
40.8%
Best Score
84.1%
High Performers (80%+)
2Performance Metrics
Max Context Window
1.1MAvg Throughput
5.7 tok/sAvg Latency
0ms+
+
+
+
All Benchmark Results for Gemini 2.5 Flash-Lite
Complete list of benchmark scores with detailed information
| FACTS Grounding | text | 0.84 | 84.1% | Self-reported | |
| Global-MMLU-Lite | text | 0.81 | 81.1% | Self-reported | |
| MMMU | multimodal | 0.73 | 72.9% | Self-reported | |
| GPQA | text | 0.65 | 64.6% | Self-reported | |
| Vibe-Eval | multimodal | 0.51 | 51.3% | Self-reported | |
| AIME 2025 | text | 0.50 | 49.8% | Self-reported | |
| LiveCodeBench | text | 0.34 | 33.7% | Self-reported | |
| SWE-Bench Verified | text | 0.32 | 31.6% | Self-reported | |
| Aider-Polyglot | text | 0.27 | 26.7% | Self-reported | |
| MRCR v2 | text | 0.17 | 16.6% | Self-reported |
Showing 1 to 10 of 13 benchmarks