Gemini 3.1 Pro Preview
GoogleGeminiProprietary
描述
Gemini 3.1 Pro is the latest model in the Gemini 3 series. It excels at complex tasks requiring broad world knowledge and advanced reasoning across modalities. Gemini 3.1 Pro uses dynamic thinking by default to reason through prompts, and features a 1 million-token input context window with 64k output tokens.
发布日期
2026-02-19
参数规模
—
上下文长度
1.0M
支持模态
audio, file, image, text, video
能力雷达图
55
general
56
coding
94
reasoning
72
science估算
80
agents
80
multimodal
Science 在缺少专门科学评测时使用推理能力代理估算。
排行榜排名
基准测试分数 (LLM Stats)
Agents
GDPval-AA
1317.00 / 3000自报
t2-bench
99.3%自报
BrowseComp
85.9%自报
MCP Atlas
69.2%自报
Terminal-Bench 2.0
68.5%自报
SWE-Bench Pro
54.2%自报
APEX-Agents
33.5%自报
Biology
GPQA
94.3%自报
SciCode
59.0%自报
Code
LiveCodeBench Pro
2887.00 / 3000自报
SWE-Bench Verified
80.6%自报
General
MMMLU
92.6%自报
MMMU-Pro
80.5%自报
MRCR v2 (8-needle)
26.3%自报
Math
Humanity's Last Exam
51.4%自报
Reasoning
ARC-AGI v2
77.1%自报
AA 评测指数
Intelligence Index57.2
Coding Index55.5
Tau21.0
Gpqa0.9
Ifbench0.8
Lcr0.7
Scicode0.6
Terminalbench Hard0.5
Hle0.4
LLM Stats 分类评分
Finance100
Legal100
General100
Code100
Reasoning100
Agents100
Language90
Search90
Spatial Reasoning80
Tool Calling80
Biology80
Chemistry80
Frontend Development80
Multimodal80
Physics80
Vision70
Math70
Long Context30
Healthcare20
定价
输入价格$2 / 1M tokens
输出价格$12 / 1M tokens
混合价格(3:1)$4.5 / 1M tokens
速度
Tokens/秒130.3 tokens/s
首Token延迟28.61s
首回答延迟28.61s
可用提供商
(LS 内部计价单位)| 提供商 | 输入价格 | 输出价格 |
|---|---|---|
| 2.5M | 15.0M |