跳轉到主要內容

GLM-4.5-Air

Z AIGLMOpen WeightMIT · Commercial OK

描述

GLM-4.5-Air is a more compact variant of GLM-4.5 designed for efficient Agentic, Reasoning, and Coding (ARC) applications. It features 106 billion total parameters with 12 billion active parameters using MoE architecture. Like GLM-4.5, it is a hybrid reasoning model providing thinking mode for complex reasoning and tool usage, and non-thinking mode for immediate responses. Despite its compact design, GLM-4.5-Air delivers competitive performance with a score of 59.8 across 12 industry-standard benchmarks, ranking 6th overall while maintaining superior efficiency. It supports 128K context length and is released under MIT open-source license allowing commercial use.

發布日期
2025-07-28
參數規模
106.0B
上下文長度
131K
支援模態
text

能力雷達圖

38
general
40
coding
79
reasoning
45
science估算
70
agents
0
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89
38.0
LS
代码能力榜159
49.0
AA
通用能力榜216
48.0
AA
数学推理73
83.0
AA
科学能力211
47.0
AA

基準測試分數 (LLM Stats)

Agents

BFCL-v376.4%自報
Terminal-Bench30.0%自報
BrowseComp21.3%自報

Biology

GPQA75.0%自報
SciCode37.3%自報

Code

LiveCodeBench70.7%自報
SWE-Bench Verified57.6%自報

Communication

TAU-bench Retail77.9%自報
TAU-bench Airline60.8%自報

Finance

MMLU-Pro81.4%自報

General

AA-Index64.8%自報

Math

MATH-50098.1%自報
AIME 202489.4%自報
Humanity's Last Exam10.6%自報

AA 評測指數

Math Index
80.7
Coding Index
23.8
Intelligence Index
23.2
Math 500
1.0
Mmlu Pro
0.8
Aime 25
0.8
Gpqa
0.7
Livecodebench
0.7
Aime
0.7
Tau2
0.5
Lcr
0.4
Ifbench
0.4
Scicode
0.3
Terminalbench Hard
0.2
Hle
0.1

LLM Stats 分類評分

Structured Output
80
Finance
80
Healthcare
80
Language
80
Legal
80
Tool Calling
70
Communication
70
General
70
Biology
60
Chemistry
60
Frontend Development
60
Math
60
Physics
60
Reasoning
60
Code
50
Agents
40
Search
20
Vision
10

定價

輸入價格$0.17 / 1M tokens
輸出價格$0.98 / 1M tokens
混合價格(3:1)$0.372 / 1M tokens

速度

Tokens/秒84.4 tokens/s
首Token延遲1.39s
首回答延遲25.10s

可用提供商

(LS 內部計價單位)

暫無提供商資料

外部連結