llmboard.ai
Benchmarks
CompareRankings
llmboard.ai
Benchmarks
CompareRankings
HomeModelsGLM 4.5 Air

Zhipu AI model product

GLM 4.5 Air

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.

Updated Aug 10, 2026. Default version: GLM-4.5-Air

Compare
LLMBoard score55.1GLM-4.5-Air
Coverage100%8 benchmark families
Context window131.1KTokens
Official input priceN/AOfficial price unavailable

On this page

  • Specification
  • Capability
  • Benchmarks
  • Arena
  • Pricing
  • Versions
  • About
  • Compare
  • Similar models
  • FAQ

Model specification

Structured fields from the published default version.

Version
GLM-4.5-Air
Released
Jul 28, 2025
Knowledge cutoff
Apr 1, 2025
Parameters
106B
Context window
131.1K
Max output
98.3K
Inputs
text
Outputs
text
Open weights
Yes
License
MIT

Capability profile

This profile uses the latest version under this unique model that has a calculated LLMBoard score. Arena and price are excluded.

GLM-4.5-Air category scores

Benchmark results

Published benchmark records for the scored version GLM-4.5-Air.

14 rows
Columns

Show columns

AA-Index0.62350.0%CAug 7, 2026
BFCL-v30.821994.4%CAug 7, 2026
TAU-bench Airline0.632390.9%CAug 7, 2026
MATH-5001.043290.3%CAug 7, 2026
TAU-bench Retail0.872575.0%CAug 7, 2026
AIME 20240.9105382.7%CAug 7, 2026
SciCode0.4161811.8%CAug 7, 2026
Terminal-Bench0.3202520.8%CAug 7, 2026
LiveCodeBench0.7217372.2%CAug 7, 2026
MMLU-Pro0.84312967.2%CAug 7, 2026
BrowseComp0.257581.8%CAug 7, 2026
Humanity's Last Exam0.1819212.1%CAug 7, 2026
SWE-Bench Verified0.68110422.3%CAug 7, 2026
GPQA0.810623354.7%CAug 7, 2026

Arena results

Preference and agent-evaluation signals from published Arena datasets.

71 rows
Columns

Show columns

textjapanese1061344.0785N/AAug 6, 2026
text factualityindustry business and management and financial operations1161395.6751N/AAug 6, 2026
text factualityindustry writing and literature and language1181374.9953N/AAug 6, 2026
text factualitymulti turn1181384.6702N/AAug 6, 2026
text factualityhard prompts english1201427.31,130N/AAug 6, 2026
textfrench1211407.6466N/AAug 6, 2026
text factualitycoding1221429.4859N/AAug 6, 2026
text factualityinstruction following1231379.01,149N/AAug 6, 2026
text style controljapanese1251327.4785N/AAug 6, 2026
textchinese1261431.21,614N/AAug 6, 2026
text factualitylonger query1261386.91,040N/AAug 6, 2026
text factualitynon english1301382.73,225N/AAug 6, 2026
text factualityexclude ties1311382.64,874N/AAug 6, 2026
text factualityoverall1321396.17,198N/AAug 6, 2026
text factualityenglish1321405.43,072N/AAug 6, 2026
text factualityindustry software and it services1331417.61,499N/AAug 6, 2026
textgerman1341370.4672N/AAug 6, 2026
textpolish1341367.41,851N/AAug 6, 2026
text factualityhard prompts1341407.03,544N/AAug 6, 2026
textindustry life and physical and social science1361407.94,924N/AAug 6, 2026
textindustry medicine and healthcare1361403.61,641N/AAug 6, 2026
textmath1361397.41,535N/AAug 6, 2026
text style controlfrench1361404.6466N/AAug 6, 2026
textspanish1381372.4723N/AAug 6, 2026
textindustry mathematical1411393.91,447N/AAug 6, 2026
textenglish1431395.014,806N/AAug 6, 2026
textoverall1441382.731,040N/AAug 6, 2026
textexclude ties1441356.321,840N/AAug 6, 2026
textindustry legal and government1451386.21,933N/AAug 6, 2026
textindustry business and management and financial operations1471371.15,505N/AAug 6, 2026
textnon english1471366.716,184N/AAug 6, 2026
textrussian1471369.31,482N/AAug 6, 2026
textindustry software and it services1481400.510,645N/AAug 6, 2026
textexpert1491368.31,388N/AAug 6, 2026
textkorean1491303.2635N/AAug 6, 2026
text style controlchinese1491412.91,614N/AAug 6, 2026
texthard prompts1501378.214,862N/AAug 6, 2026
texthard prompts english1501387.07,644N/AAug 6, 2026
textindustry writing and literature and language1501354.07,011N/AAug 6, 2026
text style controlmath1501389.31,535N/AAug 6, 2026
textcoding1511396.56,099N/AAug 6, 2026
text style controlindustry mathematical1521390.11,447N/AAug 6, 2026
textcreative writing1531341.94,010N/AAug 6, 2026
textmulti turn1531371.95,120N/AAug 6, 2026
textindustry entertainment and sports and media1541338.65,674N/AAug 6, 2026
textinstruction following1561353.88,107N/AAug 6, 2026
textlonger query1571365.16,613N/AAug 6, 2026
text style controlgerman1601348.4672N/AAug 6, 2026
text style controlspanish1601363.3723N/AAug 6, 2026
text style controlindustry life and physical and social science1611396.94,924N/AAug 6, 2026
text style controlindustry software and it services1621414.410,645N/AAug 6, 2026
text style controlcoding1631426.56,099N/AAug 6, 2026
text style controlindustry business and management and financial operations1631371.95,505N/AAug 6, 2026
text style controlenglish1641389.114,806N/AAug 6, 2026
text style controlexpert1641387.81,388N/AAug 6, 2026
text style controlindustry medicine and healthcare1641396.11,641N/AAug 6, 2026
text style controllonger query1641378.96,613N/AAug 6, 2026
text style controlhard prompts english1661402.87,644N/AAug 6, 2026
text style controlpolish1671341.21,851N/AAug 6, 2026
text style controlhard prompts1681390.614,862N/AAug 6, 2026
text style controloverall1691372.731,040N/AAug 6, 2026
text style controlinstruction following1691362.18,107N/AAug 6, 2026
text style controlmulti turn1691369.25,120N/AAug 6, 2026
text style controlkorean1701285.1635N/AAug 6, 2026
text style controlrussian1701356.91,482N/AAug 6, 2026
text style controlexclude ties1711342.921,840N/AAug 6, 2026
text style controlnon english1721352.416,184N/AAug 6, 2026
text style controlindustry legal and government1791374.31,933N/AAug 6, 2026
text style controlindustry writing and literature and language1791341.77,011N/AAug 6, 2026
text style controlcreative writing1801326.94,010N/AAug 6, 2026
text style controlindustry entertainment and sports and media1801325.05,674N/AAug 6, 2026

Pricing

Official vendor API PAYG pricing is summarized first. The table then lists individual provider offerings without treating their minimum as the official price.

Official API
N/A
Official provider
N/A
Lowest third-party
From $0.1143 input, $0.286 output per 1M via 302.AI
Tracked offerings
4
3 rows
Columns

Show columns

302.AIglm-4.5-airglobal$0.1143$0.286131.1KAug 7, 2026
LLM Gatewayglm-4.5-airglobal$0.13$0.85131KAug 7, 2026
Z.AIglm-4.5-airglobal$0.20$1.1131.1KAug 7, 2026

Official prices use only the vendor's configured official Provider and positive standard USD PAYG rates. Third-party offers remain explicitly labeled.

GLM 4.5 Air versions

All published versions linked to this unique model. The score columns identify the version used by the current overall ranking.

1 rows
Columns

Show columns

GLM-4.5-AirJul 28, 202555.1106B131.1K98.3KYesMIT

What is GLM 4.5 Air?

A concise description based on the published model registry.

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.

Use the benchmark, Arena and pricing sections above as separate evidence. A missing field means the current data snapshot does not support that claim.

Data snapshot: 2026-08-07. Editorial model content is not available in the backend.

GLM 4.5 Air vs nearby models

Open a comparison with the three ranked models immediately above and below this model.

GLM 4.5 AirvsMiniMax M1 80KGLM 4.5 AirvsMiMo V2 FlashGLM 4.5 Airvso3GLM 4.5 AirvsNova 2 LiteGLM 4.5 Airvso4 miniGLM 4.5 AirvsMiniMax M2.1

Models similar to GLM 4.5 Air

Recommendations prioritize the same model type and family, then the closest published LLMBoard score.

#48+6.9
ZA

GLM 4.5

Zhipu AI

62.0 LLMBoard

DetailsCompare
#40+9.6
ZA

GLM 5.1

Zhipu AI

64.7 LLMBoard

DetailsCompare
#38+11.0
ZA

GLM 4.7

Zhipu AI

66.2 LLMBoard

DetailsCompare
#10+25.2
ZA

GLM 5.2

Zhipu AI

80.3 LLMBoard

DetailsCompare
#66-0.1
AM

Nova 2 Lite

Amazon

55.1 LLMBoard

DetailsCompare
#67-0.2
OP

o4 mini

OpenAI

55.0 LLMBoard

DetailsCompare

FAQ

Common questions about GLM 4.5 Air.

When was GLM 4.5 Air released?

GLM 4.5 Air's default version was released on Jul 28, 2025.

How much does GLM 4.5 Air cost?

No official standard PAYG price is currently available for GLM 4.5 Air. The lowest tracked third-party offer starts at $0.1143 input and $0.286 output via 302.AI.

Who created GLM 4.5 Air?

GLM 4.5 Air is published under Zhipu AI in the model registry.

What is the context window for GLM 4.5 Air?

The default version has a 131.1K token context window.

Is GLM 4.5 Air open weight?

Yes. The default version is marked as open weight under MIT.

How many API providers offer GLM 4.5 Air?

4 published provider offerings are linked to the default version.

What models should I compare GLM 4.5 Air with?

Nearby ranked alternatives include MiniMax M1 80K, MiMo V2 Flash, o3.

Rankings

OverallCodingText ArenaPricing

Modalities

Image GenerationVideo GenerationSpeech-to-TextEmbeddings

Benchmarks

All BenchmarksReasoningMathCoding

Vendors

All VendorsOpenAIAnthropicGoogle
llmboard.aiCopyright 2026 llmboard.ai