---
title: "Cheaper alternatives to GLM 4.5 Air · Undominated.ai"
canonical: https://undominated.ai/alternatives/z-ai__glm-4.5-air/
description: "11 models score at least as high as GLM 4.5 Air and cost no more on a balanced workload. 4 give up nothing measurable; the other 7 name what they drop. Measured prices and independent scores."
---

# Cheaper alternatives to GLM 4.5 Air · Undominated.ai

> 11 models score at least as high as GLM 4.5 Air and cost no more on a balanced workload. 4 give up nothing measurable; the other 7 name what they drop. Measured prices and independent scores.

# Cheaper alternatives to GLM 4.5 Air

Z.ai · LMArena 1382.8 · $0.310/M on a balanced workload · prices as of 2026-08-28

## 4 models are both better and cheaper, giving up nothing.

Of the 133 models carrying both an independent score and a published price at the same delivery mode, **11** score at least as high as GLM 4.5 Air and cost no more under at least one workload. **4** match or beat it on every capability we hold — context, maximum output, input modes, tool use and extended reasoning. **7** do not, and every row names what it drops.

Dominance is not a property of a model. It is a property of a model, a capability lens and a workload, and all three are named on every row. Scores are LMArena Elo, used under CC BY 4.0 from the official dataset; prices are blended per million tokens. A higher score is not a drop-in replacement.

This page is built from that comparison and nothing else. When no model is both better and cheaper than GLM 4.5 Air, it is not generated.

## Where it holds

| Workload | GLM 4.5 Air $/M | Better and cheaper | With a trade |
| --- | --- | --- | --- |
| Balanced | $0.310 | 4 | 7 |
| Summarise | $0.136 | 3 | 6 |
| Chat | $0.387 | 4 | 7 |
| Code gen | $0.554 | 4 | 7 |
| Agentic | $0.175 | 3 | 6 |

Workload mixes are defined on the [methodology page](/methodology/). Cached input is priced at the cached rate, and reasoning tokens at the worse of the reasoning and output rates.

## What you would be replacing

| Intelligence | 1382.8 |
| --- | --- |
| Context window | 131K |
| Max output | 98K |
| Input modes | text |
| Tool use | yes |
| Extended reasoning | yes |

A replacement has to clear every line above, not just the score. [Full record for GLM 4.5 Air](/models/z-ai__glm-4.5-air/).

## Better and cheaper, nothing given up 4

Each of these matches or beats GLM 4.5 Air on context, maximum output, input modes, tool use and reasoning, scores at least as high, and costs no more.

| # | Model | Intelligence | $/M | Holds under |
| --- | --- | --- | --- | --- |
| 1 | Hy3 Tencent | 1441.2 +58.4 | $0.144 | Balanced −53% Summarise −43% Chat −58% Code gen −59% Agentic −53% |
| 2 | DeepSeek V4 Flash 0423 DeepSeek | 1431.6 +48.8 | $0.101 | Balanced −67% Summarise −51% Chat −76% Code gen −78% Agentic −69% |
| 3 | MiMo-V2.5 Xiaomi | 1427.3 +44.5 | $0.175 | Balanced −44% Summarise −21% Chat −60% Code gen −62% Agentic −55% |
| 4 | DeepSeek V3.2 DeepSeek | 1424.6 +41.8 | $0.290 | Balanced −6% Chat −30% Code gen −42% |

## Cheaper and higher-scoring, but you give something up 7

These score at least as high and cost no more on the two plotted axes, and lose something that is not on them. Read the last column before switching.

| Model | Intelligence | $/M | Holds under | What you give up |
| --- | --- | --- | --- | --- |
| Gemma 4 31B Google | 1441.7 +58.9 | $0.152 | Balanced −51% Summarise −33% Chat −54% Code gen −57% Agentic −41% | 98K → 16K max output |
| Gemma 4 26B A4B Google | 1434.6 +51.8 | $0.138 | Balanced −56% Summarise −39% Chat −54% Code gen −58% Agentic −37% | 98K → 16K max output |
| DeepSeek V3.2 Exp DeepSeek | 1424.4 +41.6 | $0.305 | Balanced −2% Chat −16% Code gen −36% | 98K → 66K max output |
| Qwen3 235B A22B Instruct 2507 Qwen | 1419.3 +36.5 | $0.205 | Balanced −34% Summarise −17% Chat −29% Code gen −34% Agentic −9% | 98K → 16K max output no extended reasoning |
| Step 3.5 Flash StepFun | 1403.8 +21 | $0.150 | Balanced −52% Summarise −19% Chat −53% Code gen −60% Agentic −26% | 98K → 66K max output |
| Qwen3.5-Flash Qwen | 1397.6 +14.8 | $0.114 | Balanced −63% Summarise −45% Chat −63% Code gen −67% Agentic −46% | 98K → 66K max output |
| Qwen3 30B A3B Instruct 2507 Qwen | 1384.3 tie | $0.084 | Balanced −73% Summarise −59% Chat −73% Code gen −76% Agentic −60% | 98K → 32K max output no extended reasoning |

## What this compares, and what it leaves out

 - Quality is LMArena Elo, used under CC BY 4.0 from the official dataset. The 95% confidence interval on a difference between two scores is about ±10.68 points, so a gap smaller than that is marked *tie* rather than an improvement — see [significance bands](/significance/).
 - 188 further models at this delivery mode carry a price but no independent score. They are absent from the comparison in both directions — unrated is not a zero, and an unmeasured model is neither an alternative nor a worse buy.
 - Retired models are never offered as an alternative, and a model only competes against its own delivery mode: batch trades latency for price, so it is not a like-for-like swap.
 - Nothing here measures latency, throughput, rate limits or how a model behaves on your prompts. Two models with the same index score are not interchangeable.
 - The models that nothing beats on both axes are on the [value frontier](/frontier/), and every other model something cheaper beats is [listed here](/alternatives/).