跳轉到主要內容

Gemma 4 31B (Reasoning)

GoogleGemmaOpen WeightApache 2.0 · Commercial OK

描述

Gemma 4 31B is Google DeepMind's flagship dense multimodal model with 31 billion parameters and a 256K context window. Ranks #3 among open models on Arena AI. Built from the same research as Gemini 3, it features Per-Layer Embeddings, Shared KV Cache, alternating sliding-window and global attention, and variable aspect ratio vision encoding. Achieves an estimated LMArena text score of 1452.

發布日期
2026-04-02
參數規模
30.7B
上下文長度
262K
支援模態
image, text, video

能力雷達圖

36
general
39
coding
86
reasoning
58
science估算
90
agents
70
multimodal

Science 在缺少專門科學評測時使用推理能力代理估算。

排行榜排名

領域#排名分數來源
智能体与工具1
86.0
LS
代码能力榜63
68.0
AA
通用能力榜90
69.0
AA
科学能力51
71.0
AA

基準測試分數 (LLM Stats)

Agents

t2-bench86.4%自報

Biology

GPQA84.3%自報

Finance

MMLU-Pro85.2%自報

General

MMMLU88.4%自報
LiveCodeBench v680.0%自報
MMMU-Pro76.9%自報
BIG-Bench Extra Hard74.4%自報
MRCR v266.4%自報

Healthcare

MedXpertQA61.3%自報

Math

AIME 202689.2%自報
MathVision85.6%自報
Humanity's Last Exam26.5%自報

AA 評測指數

Intelligence Index
39.2
Coding Index
38.7
Gpqa
0.9
Ifbench
0.8
Lcr
0.6
Tau2
0.6
Scicode
0.4
Terminalbench Hard
0.4
Hle
0.2

LLM Stats 分類評分

Tool Calling
90
Agents
90
Finance
90
Legal
90
Biology
80
Chemistry
80
General
80
Language
80
Physics
80
Math
70
Multimodal
70
Reasoning
70
Vision
60
Healthcare
60
Long Context
60

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34.9 tokens/s
首Token延遲1.02s
首回答延遲58.33s

可用提供商

(LS 內部計價單位)
提供商輸入價格輸出價格
Novita140K400K

外部連結