Our most intelligent AI model, built for the agentic era. Gemini 2.5 Pro leads on common benchmarks with enhanced reasoning, multimodal capabilities (text, image, video, audio input), and a 1M token context window.
发布日期2025年5月20日
参数规模—
上下文长度1.0M
许可证Proprietary
知识截止2025年1月31日
Benchmarks
评测成绩
| 评测基准 | 类别 | 分数 | 来源 |
|---|---|---|---|
| Arena AI Document Elo | arena | 94.7 | 来源 |
| MRCR | long_context reasoning general | 93.0 | 来源 |
| AIME 2024 | math reasoning | 92.0 | 来源 |
| 开源模型能力榜·IFEval | general | 90.0 | 来源 |
| 多模态开源模型榜·AI2D | vision | 89.5 | 来源 |
| 开源模型能力榜·AIME2025 | general | 88.7 | 来源 |
| Global-MMLU-Lite | general language reasoning | 88.6 | 来源 |
| 多模态开源模型榜·MMBench TEST | vision | 88.3 | 来源 |
| 多模态开源模型榜·OCRBench | vision | 86.2 | 来源 |
| 开源模型能力榜·MMLU-Pro | general | 85.8 | 来源 |
| Video-MME | multimodal vision reasoning | 84.8 | 来源 |
| 开源模型能力榜·GPQA-Diamond | general | 84.7 | 来源 |
| 多模态安全榜·Legality | security vision | 84.7 | — |
| 多模态安全榜·Health Risk | security vision | 84.6 | — |
| 多模态开源模型榜·MMVet | vision | 83.3 | 来源 |
| Arena AI Vision Elo | arena | 83.0 | 来源 |
| GPQA | reasoning general | 83.0 | 来源 |
| AIME 2025 | math reasoning | 83.0 | 来源 |
| MRCR 1M (pointwise) | long_context reasoning general | 82.9 | 来源 |
| 多模态安全榜·Misinformation | security vision | 81.5 | — |
| 多模态开源模型榜·MathVista | vision | 80.9 | 来源 |
| 多模态安全榜·Bias | security vision | 80.2 | — |
| 多模态开源模型榜 | vision | 80.1 | 来源 |
| MMMU | multimodal reasoning general | 79.6 | 来源 |
| 多模态安全榜 | security vision | 78.1 | — |
| Aider-Polyglot | general code | 76.5 | 来源 |
| LiveCodeBench v5 | reasoning general | 75.6 | 来源 |
| 多模态安全榜·Toxicity | security vision | 75.0 | — |
| 多模态开源模型榜·MMMU VAL | vision | 74.7 | 来源 |
| 多模态开源模型榜·MMStar | vision | 73.6 | 来源 |
| 开源模型能力榜 | general | 73.6 | 来源 |
| Aider-Polyglot Edit | general code | 72.7 | 来源 |
| 开源模型能力榜·LiveCodeBenchV6 | general | 71.3 | 来源 |
| Vibe-Eval | multimodal vision general | 65.6 | 来源 |
| 多模态开源模型榜·HallusionBench | vision | 64.1 | 来源 |
| SWE-Bench Verified | reasoning frontend_development code | 63.2 | 来源 |
| 多模态安全榜·Privacy | security vision | 62.5 | — |
| SimpleQA | general reasoning | 50.8 | 来源 |
| 开源模型能力榜·HLE | general | 21.1 | 来源 |
| Humanity's Last Exam | general | 17.8 | 来源 |
| ARC-AGI v2 | reasoning vision spatial_reasoning | 4.9 | 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 | $1.13 | $9.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| 302.AI | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| Abacus | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| AIHubMix | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| Auriko | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| FrogBot | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| GitHub Copilot | $1.25 | $10.00 | 128K | — | — | ✓ | ✗ | ✗ |
| $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ | |
| Vertex | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| Helicone | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| LLM Gateway | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| SAP AI Core | $1.25 | $10.00 | 1.0M | — | — | ✓ | ✗ | ✗ |
| Cortecs | $1.65 | $11.02 | 1.0M | — | — | ✓ | ✗ | ✗ |
| NanoGPT | $2.50 | $10.00 | 1.0M | — | — | ✗ | ✗ | ✗ |
价格单位：美元/百万 token，数据来自社区整理，仅供参考。