Gemini 2.5 Flash-Lite is a model developed by Google DeepMind, designed to handle various tasks including reasoning, science, mathematics, code generation, and more. It features advanced capabilities in multilingual performance and long context understanding. It is optimized for low latency use cases, supporting multimodal input with a 1 million-token context length.
发布日期2025年6月17日
参数规模—
上下文长度1.0M
许可证Creative Commons Attribution 4.0 License
知识截止2025年1月1日
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| FACTS Grounding | reasoning | 84.1 | 来源 |
| Global-MMLU-Lite | general language reasoning | 81.1 | 来源 |
| MMMU | multimodal reasoning general | 72.9 | 来源 |
| GPQA | reasoning general | 64.6 | 来源 |
| Vibe-Eval | multimodal vision general | 51.3 | 来源 |
| AIME 2025 | math reasoning | 49.8 | 来源 |
| LiveCodeBench | reasoning general code | 33.7 | 来源 |
| SWE-Bench Verified | reasoning frontend_development code | 31.6 | 来源 |
| Aider-Polyglot | general code | 26.7 | 来源 |
| MRCR v2 | long_context reasoning general | 16.6 | 来源 |
| SimpleQA | general reasoning | 10.7 | 来源 |
| Humanity's Last Exam | general | 5.1 | 来源 |
| Arc | reasoning general | 2.5 | 来源 |
Pricing
API 价格对比
| 服务商 | 输入价 | 输出价 | 上下文 | 吞吐(tok/s) | 延迟(s) | 函数调用 | 代码执行 | 联网搜索 |
|---|---|---|---|---|---|---|---|---|
| Jiekou.AI | $0.09 | $0.36 | 1.0M | — | — | ✓ | ✗ | ✗ |
| $0.10 | $0.40 | 1.0M | — | — | ✓ | ✗ | ✗ | |
| Vertex | $0.10 | $0.40 | 1.0M | — | — | ✓ | ✗ | ✗ |
| Helicone | $0.10 | $0.40 | 1.0M | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $0.10 | $0.40 | 1.0M | — | — | ✓ | ✗ | ✗ |
| NanoGPT | $0.10 | $0.40 | 1.0M | — | — | ✗ | ✗ | ✗ |
| SAP AI Core | $0.10 | $0.40 | 1.0M | — | — | ✓ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。