← Model list

GLM-4.5-Air

zhipuai·zhipuai/glm-4.5-air·GA·Open weights·glm-air series

Lighter GLM-4.5 variant for fast coding assistance and cheaper agents

At a glance
Official price$0.20 / $1.10 per 1M
Cache read $0.03 · Z.AI
Lowest paid$0.114 / $0.286
302.AI Gateway · 2× spread
Context131,072
⚠ Providers report 128,000–131,072; the table below is authoritative
Output limit98,304
Capabilities
ReasoningTool useStructured outputTemperatureAttachments
⚠ Providers report capability flags inconsistently
Modalities
Text
Knowledge cutoff2025-04
Released / updated2025-07-28 / 2025-07-28

Quality & performanceArtificial Analysis · Intelligence Index v4.1

Intelligence16.7
Value106
Coding
Agentic
Output speed77 tok/s
TTFT2.63 s
Value formulaIQ 16.7 ÷ min blended $0.157 = 106

Quality is independently evaluated by Artificial Analysis. Speed/latency are model-level medians.

Available at 16 providers15 with public prices

ProviderTierInputOutputCache readCache writeContextOutput limitStatus
302.AIGateway$0.114$0.286131,07298,304
submodel
zai-org/GLM-4.5-Air
Gateway$0.10$0.50131,072131,072
ZenMuxGateway$0.11$0.56$0.02128,00064,000
NanoGPT
zai-org/GLM-4.5-Air
Gateway$0.12$0.80$0.06128,00098,304
Hugging Face
zai-org/GLM-4.5-Air
Cloud$0.13$0.85131,07298,304
OpenRouterGateway$0.13$0.85$0.025131,07298,304
NovitaAIGateway$0.13$0.85$0.025131,07298,304
DevPass (LLM Gateway)Gateway$0.13$0.85$0.025$0131,00098,304
Kilo GatewayGateway$0.13$0.85$0.025131,07298,304
OrcaRouterGateway$0.20$1.10$0.03$0131,07298,304
Z.AIOfficialFirst-party$0.20$1.10$0.03$0131,07298,304

Sorted by blended price (input×0.75 + output×0.25) asc. The official channel always shows regardless of rank. Whether a gateway's low price is actually usable can't be verified.

Your usage cost

1302.AI$37.16
2ZenMux$39.20
3submodel$45.00
4OpenRouter$55.90
5NovitaAI$55.90
6DevPass (LLM Gateway)$55.90
12Z.AI · Official$74.60
Switch to 302.AI to save $37.44/mo (50%)
Note: this is a gateway; verify availability and rate limits yourself.

Benchmark3 items

NameConditionsScoreMetricSource
Artificial Analysis Coding Index23.8indexSource ↗
SciCode30.6percent correctSource ↗
Terminal-Bench Hard20.5success rateSource ↗

The same benchmark scores very differently across harness / dataset, so the qualifying conditions must be shown together.

Reasoning control

Toggle (on / off)effort = noneeffort = high

3 / 16 providers expose no reasoning control (reasoning_options: []).

Related models

Price historyone sample accumulated per data sync

Input listOutput listMin blended
$1.10$02026-08-052026-08-132026-08-05 · Input list $0.202026-08-06 · Input list $0.202026-08-07 · Input list $0.202026-08-08 · Input list $0.202026-08-09 · Input list $0.202026-08-10 · Input list $0.202026-08-11 · Input list $0.202026-08-12 · Input list $0.202026-08-13 · Input list $0.202026-08-05 · Output list $1.102026-08-06 · Output list $1.102026-08-07 · Output list $1.102026-08-08 · Output list $1.102026-08-09 · Output list $1.102026-08-10 · Output list $1.102026-08-11 · Output list $1.102026-08-12 · Output list $1.102026-08-13 · Output list $1.102026-08-05 · Min blended $0.1572026-08-06 · Min blended $0.1572026-08-07 · Min blended $0.1572026-08-08 · Min blended $0.1572026-08-09 · Min blended $0.1572026-08-10 · Min blended $0.1572026-08-11 · Min blended $0.1572026-08-12 · Min blended $0.1572026-08-13 · Min blended $0.157