Skip to content
Weights Not Walls
Menu

zai-org

GLM-4.5-Air

Reviewed
Size
110B parameterscounted from the safetensors header on the Hub
Licence
unknown, MIT LicenseOur reading, check the licence: You can use, change and redistribute the model for any purpose, including commercially. Keep the copyright notice and the licence text with copies. You do not have to share your changes.
On the Hub since
20 Jul 2025the date the repo was created
Context length
128k tokensfrom the model's config
Kind
Base model (no base model declared on the Hub)Glm4MoeForCausalLM
On Hugging Face
zai-org/GLM-4.5-Air1.1M downloads in the last 30 days, 650 likes

Scores

2 independent results on 2 eval sets
GSO-BenchReported

Epoch's name: GLM-4.5-Air. Best closed model on this set: claude-fable-5-1_unknown, 88.2% (Anthropic). All 42 results

HLEReported

Epoch's name: GLM-4.5-Air. Best closed model on this set: gpt-6-astra_unknown, 54.8% (OpenAI). All 56 results

Claimed by the maker Claimed

The model card has no claimed scores in a machine-readable form (the card's model-index), so none are shown. The maker may give scores in the card text: read the card.

Other views, which we link to and do not copy: look this model up on Artificial Analysis and the Arena leaderboard.

Independent scores from Epoch AI (CC-BY 4.0), fetched 19 Nov 2026: Epoch AI, ‘Capabilities & benchmarking’. Published online at epoch.ai. Retrieved from ‘https://epoch.ai/benchmarks’ [online resource]. Each row links its own source. What the badges mean.

Run it yourself

file sizes read from the Hub 19 Nov 2026

Memory needed = weights + context cache + 0.50 GB for the runtime. Context cache at 8k tokens: 1.4 GB, worked out from the model's config (zai-org/GLM-4.5-Air), at 16-bit precision. Choose a machine above to see a verdict for each format. Method

GGUF for llama.cpp, Ollama, LM Studio and others

unsloth/GLM-4.5-Air-GGUF (converted by unsloth), 145k downloads. Sizes are the real files on the Hub, all parts added up.

FormatSizeMemory neededGet it
UD-IQ3_XXS47.9 GBfile size, 2 parts49.8 GB47.9 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-IQ3_XXS
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-IQ3_XXS/GLM-4.5-Air-UD-IQ3_XXS-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q3_K_S49.0 GBfile size, 2 parts51.0 GB49.0 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q3_K_S
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q3_K_S/GLM-4.5-Air-Q3_K_S-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
UD-Q3_K_XL51.0 GBfile size, 2 parts53.0 GB51.0 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-Q3_K_XL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-Q3_K_XL/GLM-4.5-Air-UD-Q3_K_XL-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q3_K_M53.3 GBfile size, 2 parts55.2 GB53.3 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q3_K_M
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q3_K_M/GLM-4.5-Air-Q3_K_M-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
IQ4_XS56.3 GBfile size, 2 parts58.3 GB56.3 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:IQ4_XS
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "IQ4_XS/GLM-4.5-Air-IQ4_XS-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
IQ4_NL58.4 GBfile size, 2 parts60.3 GB58.4 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:IQ4_NL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "IQ4_NL/GLM-4.5-Air-IQ4_NL-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q4_058.5 GBfile size, 2 parts60.4 GB58.5 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q4_0
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q4_0/GLM-4.5-Air-Q4_0-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q4_K_S62.4 GBfile size, 2 parts64.4 GB62.4 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q4_K_S
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q4_K_S/GLM-4.5-Air-Q4_K_S-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
UD-Q4_K_XL63.1 GBfile size, 2 parts65.0 GB63.1 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-Q4_K_XL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-Q4_K_XL/GLM-4.5-Air-UD-Q4_K_XL-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q4_164.6 GBfile size, 2 parts66.6 GB64.6 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q4_1
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q4_1/GLM-4.5-Air-Q4_1-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q4_K_M68.0 GBfile size, 2 parts69.9 GB68.0 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q4_K_M
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q4_K_M/GLM-4.5-Air-Q4_K_M-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q5_K_S73.0 GBfile size, 2 parts75.0 GB73.0 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q5_K_S
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q5_K_S/GLM-4.5-Air-Q5_K_S-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
UD-Q5_K_XL77.2 GBfile size, 2 parts79.1 GB77.2 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-Q5_K_XL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-Q5_K_XL/GLM-4.5-Air-UD-Q5_K_XL-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q5_K_M77.8 GBfile size, 2 parts79.7 GB77.8 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q5_K_M
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q5_K_M/GLM-4.5-Air-Q5_K_M-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
Q6_K92.2 GBfile size, 2 parts94.1 GB92.2 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q6_K
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q6_K/GLM-4.5-Air-Q6_K-*-of-00002.gguf" --local-dir GLM-4.5-Air-GGUF
UD-Q6_K_XL94.6 GBfile size, 3 parts96.5 GB94.6 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-Q6_K_XL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-Q6_K_XL/GLM-4.5-Air-UD-Q6_K_XL-*-of-00003.gguf" --local-dir GLM-4.5-Air-GGUF
Q8_0109.4 GBfile size, 3 parts111.3 GB109.4 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:Q8_0
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "Q8_0/GLM-4.5-Air-Q8_0-*-of-00003.gguf" --local-dir GLM-4.5-Air-GGUF
UD-Q8_K_XL119.0 GBfile size, 3 parts120.9 GB119.0 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:UD-Q8_K_XL
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "UD-Q8_K_XL/GLM-4.5-Air-UD-Q8_K_XL-*-of-00003.gguf" --local-dir GLM-4.5-Air-GGUF
BF16205.8 GBfile size, 5 parts207.8 GB205.8 + 1.4 + 0.50
Commands
llama.cpp
llama-server -hf unsloth/GLM-4.5-Air-GGUF:BF16
Download
hf download unsloth/GLM-4.5-Air-GGUF --include "BF16/GLM-4.5-Air-BF16-*-of-00005.gguf" --local-dir GLM-4.5-Air-GGUF

MLX for Apple silicon

FormatSizeMemory neededRepo
MLX 4bitmlx-community/GLM-4.5-Air-4bit56.0 GBfile size57.9 GB56.0 + 1.4 + 0.50Download page
MLX 8bitlmstudio-community/GLM-4.5-Air-MLX-8bit105.8 GBfile size107.7 GB105.8 + 1.4 + 0.50Download page

Original weights size from the safetensors header, for vLLM, SGLang, Transformers

FormatSizeMemory neededRepo
BF16original weights205.8 GBfile size207.7 GB205.8 + 1.4 + 0.50
hf download zai-org/GLM-4.5-Air --local-dir GLM-4.5-Air
Other conversions (22; sizes not read)

The Hub tags 66 quantised repos as conversions of this model; we read the most downloaded and show the maker's own first.

Run it by API

Hugging Face Inference Providers, fetched 19 Nov 2026, 13:58 UTC

3 providers serve this model through the Hugging Face router. Prices are US dollars per million tokens. Precision is what the provider reports; a lower precision than the original weights can change answers. Live first, then cheapest output.

ProviderInOutContextPrecisionTokens/sFirst tokenUptime 24hToolsJSON
Novita$0.13$0.85128kbf16730.6 s100.0%yesno
SiliconFlow$0.14$0.86128kfp8531.3 s99.5%yesno
Z.AIthe maker$0.20$1.10128kfp8690.5 s100.0%yesno

Through the router, with a Hugging Face token (the router bills your Hugging Face account). Change the name after the colon to pick another provider.

curl https://router.huggingface.co/v1/chat/completions \ -H "Authorization: Bearer $HF_TOKEN" -H "Content-Type: application/json" \ -d '{"model": "zai-org/GLM-4.5-Air:novita", "messages": [{"role": "user", "content": "Hello"}]}'

Not built yet: one key for every open model, with BlockRun, and our own tested-quality score per provider. Until then, precision and measured speed are the best guide to a provider here.

Planned for this page Not built yet

Try it in private chat
A link to Solana AI's confidential chat for the models it serves. This model is not listed as served there.
Fine-tune it
Fine-tuning a few base models from this page and publishing the result here.
Creators and earnings
Pages for people who publish fine-tunes and adapters of this model, and what they earn.
Fund and bounties
Bounties for new open models, paid from a public fund. How the fund will work.
Contribute compute
How to give compute to open model work, with Akash.

Model data from the Hugging Face Hub 19 Nov 2026, 13:43 UTC. The repo. Found a mistake on this page? Tell us.