← 全球大模型库
Gemini 3.1 Pro Preview
Google · google/gemini-3.1-pro-preview
价格 $2 起 ~ $4 / 百万 tokens · 共 17 家算力服务商部署
各服务商对比
服务商模型 ID上下文最大输出输入 $/M输出 $/M缓存 $/MAPI 端点
Vercel AI Gateway
google/gemini-3.1-pro-preview
1M
64K
$2
$12
$0.2
—
Google
gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
—
ZenMux
google/gemini-3.1-pro-preview
1M
64K
$2
$12
$0.2
zenmux.ai ↗
Merge Gateway
google/gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
—
Kilo Gateway
google/gemini-3.1-pro-preview
1M
65.5K
$2
$12
—
api.kilo.ai ↗
OpenRouter
google/gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
openrouter.ai ↗
AIHubMix
gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
—
Vertex
gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
—
NanoGPT
google/gemini-3.1-pro-preview
1M
65.5K
$2
$12
$0.2
nano-gpt.com ↗
模型参数
上下文窗口1M
最大输出65.5K
知识截止2025-01
发布日期2026-02-19
系列gemini-pro
最后更新2026-02-19
输入模态文本图像视频音频PDF
输出模态文本
能力
推理工具调用结构化输出温度调节文件附件
基准跑分
SWE-Bench Pro 54.2 (resolve rate) · Terminal-Bench 70.3 (success rate) · SWE-Bench Pro 46.1 (resolve rate) · SWE-Atlas Codebase QnA 13.5 (score) · SWE-Atlas Refactoring 33.81 (score) · SWE-Atlas Test Writing 29.84 (score) · Artificial Analysis Coding Agent Index 43 (average pass@1) · SWE-Atlas Codebase QnA 45.6 (pass@1) · SWE-Bench Pro 15.1 (pass@1) · Terminal-Bench 68.3 (pass@1) · GPQA Diamond 94.3 (accuracy) · Humanity's Last Exam 44.4 (accuracy) · ARC-AGI-2 77.1 (accuracy) · MMMU Pro 80.5 (accuracy) · MCP Atlas 78.2 (success rate) · OSWorld-Verified 76.2 (success rate) · CharXiv Reasoning 83.3 (accuracy) · GDPval-AA 1314 (Elo)
Google 的其他模型
百灵鸟（EchoBird）