Tier B — Production
Runs in:CNMade in:China
Z.ai (GLM / Zhipu)

GLM-4.5 Air

Tier B — Production · 131K tokens

Tokonomix Editorial Team·Reviewed by Mes Kalkan··

GLM-4.5 Air is the lightweight sibling of GLM-4.5 — a smaller, cheaper model tuned for speed and volume. It is the GLM you reach for when cost-per-token and throughput matter more than peak quality.

Section 01

Speed analysis

Latency measured across all benchmark runs. P50 (median) and P95 (95th percentile) give a realistic picture of response speed under normal and peak load.

P50 latency (median)P95 latency56 runs
369815215935237173150007-0807-22ms
Section 02

Quality scores

Evaluation results from judge-model scoring across diverse task categories. Scores reflect coherence, accuracy and instruction-following.

100
Coding
45
Creative
Section 03

Pricing history

Direct provider rates per million tokens, plus a typical-conversation cost estimate.

💰
API rates — GLM-4.5 Air
$0.2000 per 1M input tokens
$1.10 per 1M output tokens
≈ $0.0003 per typical conversation (800 tokens)
Input vs output price (per 1M tokens)
per 1M input tokens$0.2000
per 1M output tokens$1.10

Pricing over time

Input & output per 1M tokens · step-line = price changes

$0.2000

input / 1M

— stable

$1.10

output / 1M

— stable

2026-07-122026-07-192026-07-19
Input
Output
Price change
⟳ synced weekly
Section 04

Tokens per second

Throughput in tokens per second, derived from measured P50 latency. Higher is better; fluctuations track provider-side load.

Throughput (tokens / s)52 / avg 47
5366

Estimated from P50 latency × 200 output tokens — the absolute number depends on this assumption; the trend is what matters.

Section 05

Capabilities

jsonnotes: GLM emits a non-standard reasoning_content field beside content; read content for the answer.toolsreasoning
Section 06

Availability

Availability

How often this model answers when we call it — measured across real API requests and live tests over the last 30 days. This is separate from quality: these numbers only tell you whether the model responds, not how good the answer is.

Last 7 days

—

Last 30 days

100.0%

n=1

Median response time

5,107ms

n=1

Based on 175 measurements over the last 30 days.

Technical details

Only live API calls and live-test requests count — internal probes and benchmark runs are excluded.

Calls with a custom API key (BYOK) are excluded: those failures are key-specific, not a sign of model downtime.

Failed calls are NOT included in quality scores — quality is measured on successful responses only. Availability and quality are independent signals.

Median response time (p50) across successful calls with a recorded duration. Outliers (very slow or very fast calls) pull the median less than the average.

Total calls (30d)

1

OK responses (30d)

1

Total calls (7d)

0

OK responses (7d)

0

Section 07

Tokonomix benchmark verdicts

⚖️
Endorsed by 1 judge
Independent LLM judges evaluated this model on our weekly intelligence tests
claude-sonnet-4-574/100 · 6 runs
4 correct0 partial2 wrong67% accuracy
● 2026-07-19

GLM-4.5 Air adds tool support, reasoning; core performance unchanged

GLM-4.5 Air has expanded its capabilities with the addition of JSON mode, tool calling, and reasoning features, marking a significant enhancement to its functional toolkit. However, performance metrics across core tasks remain largely consistent with the previous benchmark window. The model continues to demonstrate strong competency in coding tasks while showing weaknesses in creative writing domains. The addition of tool support positions GLM-4.5 Air as more versatile for enterprise and application development scenarios, enabling function calling and structured output generation that were previously unavailable. The reasoning capability suggests potential for more complex problem-solving workflows, though benchmark scores indicate the fundamental quality of outputs has not shifted dramatically. Users evaluating this model should note that while the feature set has grown substantially, making it suitable for a broader range of use cases, the underlying performance characteristics remain steady. Creative writing tasks still represent a relative weakness compared to technical applications. The expanded capabilities make GLM-4.5 Air a more complete offering for developers seeking structured interactions and tool integration, while those prioritizing creative or narrative generation may still find limitations.

Quality

—

Latency p50

—

Test runs

0

✓ JSON and tool support added✓ New reasoning capability introduced✗ Creative writing remains weak✗ Core performance metrics unchanged
Section 08

Full model profile

GLM-4.5 Air: the lightweight, low-cost GLM

GLM-4.5 Air is the lightweight sibling of GLM-4.5 — a smaller, cheaper model tuned for speed and volume. It is the GLM you reach for when cost-per-token and throughput matter more than peak quality.

z.ai publishes GLM-4.5 Air at $0.20 per 1M input tokens and $1.10 per 1M output tokens — the cheapest paid GLM tier, well below the standard GLM-4.5.

It advertises a 128K-token context window.

Architecture & training signals

GLM-4.5 Air is the smaller, efficiency-oriented variant of GLM-4.5 from Zhipu AI. It keeps tool-calling and JSON support in a lighter package aimed at throughput and cost, over an OpenAI-compatible endpoint. Like the rest of the GLM line, it returns a non-standard reasoning_content field alongside content in its OpenAI-compatible responses; integrations should read content for the final answer and treat reasoning_content as an optional trace.

Where it shines

  • High-volume, cost-sensitive workloads — classification, extraction, bulk drafting.
  • Lower latency than the full GLM-4.5.
  • The lowest paid GLM price point.

Where it falls short

  • Lower ceiling than the full-size models — not for the hardest reasoning.
  • No Tokonomix benchmark data yet; validate on your tasks.
  • Non-EU hosting.

Real-world use cases

  • Bulk data processing where cost dominates.
  • Interactive features needing fast, cheap responses.
  • First-pass drafting before a stronger model refines.

Tokonomix benchmark snapshot

GLM-4.5 Air is newly registered on Tokonomix and not yet activated, so we have not run it through our weekly intelligence test or speed benchmark. There are no Tokonomix scores to report yet — and we will not invent any.

When it goes live, it enters the same weekly harness as every other model: identical prompts, an independent cross-family judge, and reproducible latency and cost measurements. Until then, treat the pricing and capability notes on this page as the vendor-published starting point, not as measured Tokonomix results.

EU privacy & data residency

GLM-4.5 Air is built by Zhipu AI (z.ai), a China-headquartered lab, and is served from non-EU infrastructure. This is important to state plainly: routing a prompt to this model is not an EU-data-residency or GDPR-sovereign choice, and Tokonomix will never tag it as one.

If your use case requires data to stay within the EU, pick a model whose provider is EU-hosted (for example our OVH or Azure-EU routes) rather than a GLM model. Tokonomix keeps z.ai out of every EU-only / sovereign routing set by design. Use GLM where its capability or price is the priority and cross-border processing is acceptable for that workload.

Verdict & alternatives

GLM-4.5 Air is the GLM economy tier: cheap and quick, with a sensibly lower ceiling. When cost matters even more, the GLM-4.7/4.5 Flash tiers are free; when quality matters more, step up to full GLM-4.5 or above.

Last automated test
Jul 22, 2026 · 02:00 UTC · Speed benchmark
P50 latency
3849 ms
P95 latency
4999 ms
Errors
0 / 6 runs
Last reviewed by Tokonomix Team·July 8, 2026