跳轉到主要內容

Gemma 3n E2B Instruct

GoogleGemma開源權重Gemma · 商用許可

描述

Gemma 3n is a generative AI model optimized for use in everyday devices, such as phones, laptops, and tablets. It features innovations like Per-Layer Embedding (PLE) parameter caching and a MatFormer model architecture for reduced compute and memory. These models handle audio, text, and visual data, though this E4B preview currently supports text and vision input. Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models, and is licensed for responsible commercial use.

發布日期
2025-06-26
參數規模
1.9B
上下文長度
—
支援模態
—

能力雷達圖

40
general
30
coding
40
reasoning
17
science估算
37
agents
0
multimodal

缺少專門科學評測時,Science 由 LLM Stats 科學得分或推理能力估算。

排行榜排名

暫無排名資料

基準測試分數 (LLM Stats)

(LLM Stats (zeroeval))

Code

Codegolf v2.211.0%自報

General

TriviaQA60.8%自報
MMLU60.1%自報
Global-MMLU-Lite59.0%自報
Global-MMLU55.1%自報
Include38.6%自報
OpenAI MMLU22.3%自報

Language

BoolQ76.4%自報
WMT24++42.7%自報
MMLU-Pro40.5%自報
MMLU-ProX8.1%自報
ECLeKTic2.5%自報

Math

MGSM53.1%自報
HiddenMath27.7%自報
AIME 20256.7%自報

Reasoning

PIQA78.9%自報
ARC-E75.8%自報
HellaSwagAI2 (2019)72.2%自報
Winogrande66.8%自報
HumanEvalOpenAI (2021)66.5%自報
MBPP0.57 / 100自報
DROP53.9%自報
ARC-C51.7%自報
Social IQa48.8%自報
BIG-Bench Hard44.3%自報
GPQANYU + Cohere + Anthropic (2023)24.8%自報
LiveCodeBench v518.6%自報
Natural Questions15.5%自報
LiveCodeBench13.2%自報

AA 評測指數

(Artificial Analysis)

暫無 AA 評測資料

LLM Stats 分類評分

(LLM Stats (zeroeval))
Language
50
Physics
50
Psychology
50
Reasoning
50
Creativity
50
Legal
40
Math
40
Finance
40
General
40
Healthcare
40
Code
30
Search
20
Biology
20
Chemistry
20

定價

輸入價格免費
輸出價格免費

速度

暫無速度資料

供應商價格排行

暫無提供商資料

外部連結