Gemini 2.0 Flash-Lite (Feb '25)
GoogleGeminiProprietary
描述
A Gemini 2.0 Flash model optimized for cost efficiency and low latency
發布日期
2025-02-25
參數規模
—
上下文長度
—
支援模態
image, text
能力雷達圖
26
general
19
coding
51
reasoning
38
science
41
agents
80
multimodal
排行榜排名
基準測試分數 (LLM Stats)
(LLM Stats (zeroeval))Audio
CoVoST2
38.4%自報
Factuality
SimpleQA
21.7%自報
General
Global-MMLU-Lite
78.2%自報
Language
MMLU-Pro
71.6%自報
Long Context
MRCR 1M
58.0%自報
Math
MATH
86.8%自報
HiddenMath
55.3%自報
Multimodal
MMMU
68.0%自報
Reasoning
FACTS Grounding
83.6%自報
Bird-SQL (dev)
57.4%自報
GPQANYU + Cohere + Anthropic (2023)
51.5%自報
LiveCodeBench v5
28.9%自報
Vision
EgoSchema
67.2%自報
AA 評測指數
(Artificial Analysis)Math 500(OpenAI (2024), subset of Hendrycks et al. MATH (2021))87.3
Mmlu Pro(TIGER-Lab (Univ. of Waterloo, Toronto, CMU, 2024))72.4
Gpqa(NYU + Cohere + Anthropic (2023))53.5
Aime(MAA (Mathematical Association of America))27.7
Livecodebench(UC Berkeley + MIT + Cornell (2024))18.5
Intelligence Index(Artificial Analysis)7.4
Hle(Center for AI Safety + Scale AI (2025))3.4
LLM Stats 分類評分
(LLM Stats (zeroeval))Grounding80
Legal70
Math70
Multimodal70
Finance70
Healthcare70
Vision70
Language60
Long Context60
Reasoning60
General60
Physics50
Factuality50
Biology50
Chemistry50
Speech To Text40
Audio40
定價
輸入價格免費
輸出價格免費
混合價格(3:1)免費
速度
Tokens/秒0.0
首Token延遲0.00s
首回答延遲0.00s
供應商價格排行
暫無提供商資料