Model comparison · accepted catalogue facts 2026-11-22

GLM 4.5 Air vs GLM 4.5V

GLM 4.5 Air has an equal-or-higher measured score and an equal-or-lower estimated price. Switching from GLM 4.5V loses capability or fails a requirement. Review the migration checks. The reported score intervals do not overlap; this is not a task-specific success guarantee.

Both models belong to the Z.ai GLM family. Inclusion does not recommend either model. This is a direct pair comparison; global frontier membership is not its verdict.

GLM 4.5 Air

LMArena general · Elo · higher is better

Scale starts at 1040 Elo

1383.7

Balanced effective USD / 1M tokens · lower is better

$0.310/M

GLM 4.5V

LMArena general · Elo · higher is better

Scale starts at 1040 Elo

1332.6

Balanced effective USD / 1M tokens · lower is better

$0.900/M

In-Depth Analysis

GLM 4.5 Air vs GLM 4.5V: Measured Technical Breakdown

Comprehensive comparison across measured human-preference evaluation, effective production token costs, architectural limits, and migration trade-offs.

Measured Capability (LMArena)
GLM 4.5 Air leads by 51.1 Elo

GLM 4.5 Air: 1383.7 ±4.23 · GLM 4.5V: 1332.6 ±8.43 Statistically significant preference separation

Cost Comparison Across Canonical Production Workloads

Rates represent effective cost per 1M tokens under standard token blends. Cached prompt reads are factored according to workload context reuse.

Workload & RatioGLM 4.5 AirGLM 4.5VCost DifferenceAdvantage
Balanced 3 tokens in per 1 out$0.310 /1M tokens$0.900 /1M tokens$0.590 (66% diff)GLM 4.5 Air costs less
Summarise Huge input, tiny output$0.136 /1M tokens$0.520 /1M tokens$0.384 (74% diff)GLM 4.5 Air costs less
Chat Long cached system prompt$0.387 /1M tokens$0.933 /1M tokens$0.547 (59% diff)GLM 4.5 Air costs less
Code gen Output dominates$0.554 /1M tokens$1.28 /1M tokens$0.727 (57% diff)GLM 4.5 Air costs less
Agentic Context replayed each turn$0.176 /1M tokens$0.488 /1M tokens$0.313 (64% diff)GLM 4.5 Air costs less

Architectural & Operational Specifications

SpecificationGLM 4.5 AirGLM 4.5VOperational Note
Context window131,072 tokens65,536 tokensGLM 4.5 Air has a 2.0× larger context window
Max output generation98,304 tokens16,384 tokensGLM 4.5 Air allows longer single-turn completions
Base input price$0.130 / 1M tokens$0.600 / 1M tokensGLM 4.5 Air input rate is cheaper
Base output price$0.850 / 1M tokens$1.80 / 1M tokensGLM 4.5 Air output rate is cheaper
Cached input price$0.025 / 1M tokens$0.110 / 1M tokensGLM 4.5 Air offers lower cached prompt cost
Reasoning & thinking tokensSupported; standard ratesSupported; standard rates—
Input modalitiestexttext, image—
Tool calling / function useSupportedSupported—
Open weights & licenceProprietary (API-only)Proprietary (API-only)—
Provider & reseller marketZ.ai · 3 provider offers (1.54× spread)Z.ai · 2 provider offers—

When to Choose GLM 4.5 Air

  • Higher measured general capability (1383.7 Elo on human evaluation).
  • Larger context window (131,072 tokens) for long-document or repo-level context.
  • Higher maximum completion limit (98,304 tokens).
  • Direct integration with Z.ai tooling and API features.
Risks when moving from GLM 4.5V → GLM 4.5 Air:
  • image input is not supported by the candidate.

When to Choose GLM 4.5V

  • Broader multimodal input support: text, image.
  • Direct integration with Z.ai tooling and API features.
Risks when moving from GLM 4.5 Air → GLM 4.5V:
  • Context window: 131,072 → 65,536 tokens.
  • Output limit: 98,304 → 16,384 tokens.

Compare · model decision

Choose two or three models

Inspect the differences, estimate your workload and keep the evidence. All standard catalogue models are available, including unrated ones.

Catalogue facts
2026-11-22

2 / 3 selected · 0 requirements
Open public comparison link
Public workload and requirements

These settings are included in the share link. The fixed-request scenario and private evaluation data below are excluded.

A ↔ B

GLM 4.5 Air has an equal-or-higher measured score and an equal-or-lower estimated price. Switching from GLM 4.5V loses capability or fails a requirement. Review the migration checks. The reported score intervals do not overlap; this is not a task-specific success guarantee.

Estimated blend: 75% input, 25% output; 0% cached input. Base context tier. Prompt length selects the billing tier. An unlisted cache-read rate uses that tier’s full input rate in this blend; missing required input or output rates prevent an estimate. The fixed-request estimate below validates actual context and output limits and requires a published cache-read rate when cache hits are assumed.

What changes between these models

Quality first, then estimated price. LMArena scores are human-preference measurements, not a benchmark of your application. A missing score is unrated.

  • GLM 4.5 Air · Meets the stated model requirements
  • GLM 4.5V · Meets the stated model requirements
Material differences and unknown evidence. Deltas compare each model with model A. Unknown facts remain visible.
Measured factA GLM 4.5 AirB GLM 4.5V
LMArena general score (Elo)1,383.71,332.6Δ vs A: -51.1 Elo
Score interval half-width (± Elo)4.238.43Δ vs A: +4.2 Elo
LMArena document score (Elo)UnknownUnknown
Estimated effective USD / 1M tokens$0.31$0.9Δ vs A: +0.59 USD/M
Base input USD / 1M tokens$0.13$0.6Δ vs A: +0.47 USD/M
Base output USD / 1M tokens$0.85$1.8Δ vs A: +0.95 USD/M
Base cached-input USD / 1M tokens$0.025$0.11Δ vs A: +0.085 USD/M
Context window (tokens)131,07265,536Δ vs A: -65,536 tokens
Output limit (tokens)98,30416,384Δ vs A: -81,920 tokens
Input modalitiestexttext, image
Open weightsUnknownUnknown
Recorded licenceUnknownUnknown
Distinct sellers in the catalogue32Δ vs A: -1
Serving precision differs across offersSupportedNot supportedΔ vs A: -1
Retirement evidenceNone announced; not a guaranteeretirement date recorded
Recorded retirement dateUnknown2026-12-31
4 shared measured facts
Shared measured facts. Deltas compare each model with model A. Unknown facts remain visible.
Measured factA GLM 4.5 AirB GLM 4.5V
Benchmark reasoning effortdefaultdefault
Tool useSupportedSupportedΔ vs A: 0
ReasoningSupportedSupportedΔ vs A: 0
Model providerZ.aiZ.ai

Provider counts and precision flags describe catalogue offers. They do not establish the precision, availability, region, tool behaviour or limits of your chosen endpoint. Open each model’s provider table before switching.

Complete price ladders and billing conditions

GLM 4.5 Air

Rates below are USD per 1M tokens. A threshold selects that whole input/output rate pair.

Input-token thresholdInputOutputCached input
Base, up to and including first threshold$0.13$0.85$0.025

GLM 4.5V

Rates below are USD per 1M tokens. A threshold selects that whole input/output rate pair.

Input-token thresholdInputOutputCached input
Base, up to and including first threshold$0.6$1.8$0.11
Migration preflight · compare each candidate with model A

Keeping GLM 4.5 Air is a valid decision. No pairwise comparison certifies a drop-in replacement.

GLM 4.5 Air → GLM 4.5V

Known loss or failed requirement

  • Context window: 131,072 → 65,536 tokens.
  • Output limit: 98,304 → 16,384 tokens.
  • Provider endpoint, serving precision, availability and task success need your own verification. Catalogue model capabilities do not certify an endpoint.
Check alternatives to GLM 4.5 Air

Estimate a fixed request shape

Private to this tab unless you save locally or explicitly export private details. These are estimates, not observed invoices. Cross-model token counts and task success are assumptions.

0.1 extra attempts means 10 extra full attempts per 100 requests. Cache writes, tool charges, taxes and unlisted charges are excluded.

  • GLM 4.5 AirEstimated total: $0.685 1,000 full attempts · base rates
  • GLM 4.5VEstimated total: $2.1 1,000 full attempts · base rates
Prompt and output sensitivity · sampled scenarios

Every point recalculates the entire request shape. Tier-boundary samples include one token below, at and above each threshold. Ordering between samples is not guaranteed. Lowest known cost excludes unknown estimates and is not a replacement recommendation.

Vary input; hold output, requests, cache and retries fixed.
Input tokensGLM 4.5 AirGLM 4.5VCoverage
2,000$0.685Lowest known$2.1All computed
4,096$0.95748Lowest known$3.3576All computed
32,768$4.68484Lowest known$20.5608All computed
131,072ineligibleInput plus output exceeds the published context window.ineligibleInput plus output exceeds the published context window.Incomplete
Vary output; hold input, requests, cache and retries fixed.
Output tokensGLM 4.5 AirGLM 4.5V
128$0.3688$1.4304
2,048$2.0008$4.8864
8,192$7.2232$15.9456

Bring your observed evaluation results

JSON is parsed locally. Prompts, filenames, individual responses and extra fields are not retained. Import only aggregate measurements for selected models. Task/version labels do not establish that two evaluations used the same dataset.

Expected format
[{"model":"provider/exact-model-id","task":"support-answer","version":"rubric-1","attempts":100,"successes":84,"costUsd":2.4,"latencyMs":950}]

Use the exact catalogue ID. Replace the example values with observed measurements. Omit costUsd or latencyMs when they are unknown. No public score is inferred from these observations.

Keep the decision and its evidence

A portable record retains selected facts, complete tier ladders, public requirements and the publication manifest. It also embeds the full permitted public catalogue to verify the input artifact checksum. Files are larger for this reason (up to 2 MB); local storage is limited to 12 records. Team review is local to this browser; there is no hosted collaboration.

Saving is opt-in and includes private review fields. Shared comparison links never include them. Version 1 Check references remain separate and are not overwritten or guessed into this format.

Saved local decisions · 0

No version 2 decision records are stored in this browser.

Questions this page answers

Is GLM 4.5 Air a better-and-cheaper replacement for GLM 4.5V?

GLM 4.5 Air has an equal-or-higher measured score and an equal-or-lower estimated price. Switching from GLM 4.5V loses capability or fails a requirement. Review the migration checks. The reported score intervals do not overlap; this is not a task-specific success guarantee.

Which model is higher rated: GLM 4.5 Air or GLM 4.5V?

GLM 4.5 Air scores higher on LMArena general (1383.7 ±4.2) compared to GLM 4.5V (1332.6 ±8.4), a margin of 51.1 Elo points. The measured confidence intervals do not overlap, demonstrating a statistically significant preference lead on human evaluations.

Which model is more cost-effective: GLM 4.5 Air or GLM 4.5V?

On a standard balanced workload (3 prompt tokens per 1 generated token), GLM 4.5 Air costs $0.310 per 1M tokens versus $0.900 for GLM 4.5V. GLM 4.5V is approximately 2.9× the price of GLM 4.5 Air. Base input rates are $0.130 for GLM 4.5 Air and $0.600 for GLM 4.5V; base output rates are $0.850 for GLM 4.5 Air and $1.80 for GLM 4.5V.

How do their context windows and generation limits compare?

GLM 4.5 Air provides a 131,072-token context window with a maximum output limit of 98,304 tokens. GLM 4.5V provides a 65,536-token context window with a maximum output of 16,384 tokens. GLM 4.5 Air offers a 2.0× larger context capacity, making it suitable for larger repositories and multi-document workflows.

What capabilities are lost when switching between GLM 4.5 Air and GLM 4.5V?

Switching from GLM 4.5 Air to GLM 4.5V involves: Context window: 131,072 → 65,536 tokens; Output limit: 98,304 → 16,384 tokens. Conversely, moving from GLM 4.5V to GLM 4.5 Air involves: image input is not supported by the candidate. Run test suites on your specific prompt templates and tool declarations before cutting over production traffic.

Can GLM 4.5 Air or GLM 4.5V be self-hosted or deployed privately?

Both GLM 4.5 Air and GLM 4.5V are proprietary cloud models offered via managed API endpoints. GLM 4.5 Air is served by Z.ai and 3 catalogue offers, while GLM 4.5V is served by Z.ai and 2 catalogue offers.

What is included in this comparison?

Accepted model capabilities, context and output limits, complete price-tier ladders, independent LMArena measurements and missing evidence. Provider endpoint precision and application success still need verification.

Can I compare a different workload or keep a decision record?

Yes. Change the public workload, prompt length and requirements in the comparison workspace. Fixed-request scenarios, aggregate evaluations and review notes stay local unless explicitly exported.

Evidence & Ask