alt.freeweights.glm / GLM-4-32B-Base-32K

From: arcee-ai <huggingface.co/arcee-ai>Newsgroups: alt.freeweights.glmSubject:

GLM-4-32B-Base-32K

Date: 2026-11-15Commit: af83a73Signed: 2DC28CC7B26A283B ✓

Text generation model with 32.6B parameters. Safetensors weights, 65.1 GB in all. Published by arcee-ai under MIT.

32.6B params65.1GBGLM-4-32B-Base-32Ksafetensorspermissive

FactsParams32.6BSize65.1 GBFormatssafetensorsTasktext generationLicenseMITFrom sourcepermissive

$ fw get arcee-ai/GLM-4-32B-Base-32K
# manifest sha256 ef44ca2be1a3… · 65.1 GB · 22 files
$ fw verify ./GLM-4-32B-Base-32K arcee-ai/GLM-4-32B-Base-32K

From sourceHosted (signed)arcee-ai/GLM-4-32B-Base-32KLast check: 2026-11-15 04:15 UTC

Run at home Get help

Get help: custom fine-tune or private deploy. We review your setup before we quote.

Works with

Run24 Fine-tune29 Convert15 Evaluate13 Scan13 Apps58

Runs with

Ranked by fit to this model (safetensors), then by GitHub stars.

vLLM exo Text Generation WebUI

SGLang MLC LLM KTransformers LMCache Xinference LMDeploy mistral.rs Mooncake FlashInfer

Show all 24

AIBrix fastllm ExLlamaV3

Transformers LocalAI tinygrad gpu-unlock Triton Inference Server BentoML CTranslate2 ExLlamaV2 LitServe

Convert with

mergekit TensorRT Model Optimizer

LLM Compressor Neural Compressor AutoRound AQLM GPTQModel ONNX Core ML Tools ExecuTorch

Show all 15

Evaluate with

Opik DeepEval Ragas lm-evaluation-harness OpenCompass SWE-bench

simple-evals TruLens HELM

LightEval

Evaluate Weave

Show all 13

Autoevals

Scan with

promptfoo garak Guardrails NeMo Guardrails Giskard DeepTeam AI Fairness 360 Fairlearn Detoxify MCP Scanner AgentDojo ModelScan

Show all 13

picklescan

Overview Model card Files Use it License Provenance Model tree

Overview

Size: 65.1 GB
Weights: 65.1 GB (safetensors)
Parameters: 32.6B
Formats: safetensors
Task: text generation
License: MIT (permissive)
Files: 22, each with a sha256
Publisher: arcee-ai
Added: 2026-11-15
Section: alt.freeweights.glm

Lineage

Original release from the publisher.

Base model: zai-org/GLM-4-32B-Base-0414

Where these bytes come from

Publisher at the pinned commit→You

Sources are tried in order. A file counts only when its size and sha256 match the manifest. Sources are checked when you ask for them, not on a timer.

Sources

Badge	Host	Last check	Result
From source	huggingface.co	2026-11-15 04:15 UTC	passed

Model card

From the publisher at commit af83a73. Shown as plain text with links; images open at the publisher.

GLM-4-32B-Base-32K#

GLM-4-32B-Base-32K is an enhanced version of THUDM's GLM-4-32B-Base-0414, specifically engineered to offer robust performance over an extended context window. While the original model's capabilities degraded after 8,192 tokens, this version maintains strong performance up to a 32,000-token context, making it ideal for tasks requiring long-context understanding and processing.

This model was developed as a proof-of-concept to validate that a merging-centric approach to context extension can be successfully applied to larger-scale models. The techniques employed resulted in an approximate 5% overall improvement on standard base model benchmarks while significantly improving 32k recall.

More details can be found in our blog post here where we applied this work to our upcoming AFM 4.5B

Model Details#

Architecture Base: THUDM/GLM-4-32B-Base-0414
Parameter Count: 32B
License: MIT

Improvements#

The primary improvement in this model is its enhanced long-context capability. The following methods were used to achieve this:

Targeted Long-Context Training: The model underwent continued pretraining on sequences up to its full 32,000 token context length.
Iterative Merging: Various model checkpoints were iteratively merged to combine the benefits of different training runs, enhancing both long-context and short-context performance.
Short-Context Distillation: Knowledge from the original high-performing short-context model was distilled into the long-context-trained model to recover and retain its initial capabilities on shorter tasks.

As a result, where the original model's performance on the Needle in a Haystack (NIAH) benchmark would decline after 8,000 tokens, this extended version maintains reliable performance across the entire 32,000 token context window.

Benchmarks#

Benchmark	GLM-4-32B-Base-0414	GLM-4-32B-Base-32K
arc_challenge	59.39%	64.93%
arc_easy	85.44%	87.88%
hellaswag	64.75%	65.40%
mmlu	77.05%	77.87%
piqa	81.61%	83.19%
truthfulqa_mc2	49.27%	50.07%
winogrande	78.69%	80.03%

NIAH Benchmark Results Comparison#

Model	Task	4,096	8,192	16,384	24,576	32,768
GLM-4-32B-Base-0414
	niah_single_1	100.0%	100.0%	77.0%	5.2%	1.2%
	niah_single_2	100.0%	100.0%	73.4%	2.6%	0.0%
	niah_single_3	100.0%	99.8%	48.0%	1.4%	0.0%
GLM-4-32B-Base-32k
	niah_single_1	100.0%	100.0%	100.0%	99.2%	99.6%
	niah_single_2	100.0%	100.0%	99.2%	80.2%	68.8%
	niah_single_3	100.0%	99.6%	95.6%	86.6%	61.0%

NIAH Averages#

Model	4,096	8,192	16,384	24,576	32,768
GLM-4-32B-Base-0414	100.0%	99.9%	66.1%	3.1%	0.4%
GLM-4-32B-Base-32k	100.0%	99.9%	98.3%	88.7%	76.5%

Use Cases#

This model serves as a new base for continued training at 32K context

License#

GLM-4-32B-Base-32K (32B) is released under the MIT license following with the original model's license.

If you have questions or would like to share your experiences using GLM-4-32B-Base-32K (32B), please connect with us on social media. We’re excited to see what you build—and how this model helps you innovate!

Files

22 files · 65.1 GBWeights 65.1 GB (safetensors). Sizes and sha256 from the signed manifest.

Show all 22 files with their sha256

File	Size	sha256
.gitattributes	1.63 KB	9b8c6f9c88b4c8978a9425ffe9383daf5d0ea89e5c47c534fe4af9326e5a9fca
README.md	3.99 KB	4df3520370d0012a06435b56425b818b208fc9fb095bb9242ed247e055ff1d01
config.json	652 B	210e8c173d6d4ec79f8644556766890defc1aa657da742f146eff44457e62da2
context_benchmarks.png	152 KB	5ebc1f5b1d15141919ad1811f888eb027e8f739c7c9c3bd056c3fc49468a8d81
model-00001-of-00014.safetensors	4.93 GB	91303db34a6d8e7b82912f7dd89cf7ae9a99035758b8532d34840d61c2fb538b
model-00002-of-00014.safetensors	4.95 GB	1b139c5ca1d95662fcadc39a080a8d527cbf0ebe07b706265685112647113844
model-00003-of-00014.safetensors	4.98 GB	fb8feb1140ca9aa23a5cbce7d1a4f41ef17be54857d131084b79fb93e578e396
model-00004-of-00014.safetensors	4.96 GB	01c6388f05abc52efbe35eac6e2bb6390c977710d09e2924832bc1a2c0cef66f
model-00005-of-00014.safetensors	4.95 GB	1b7501a4795bc57da1c60deff0e845308c15ab915506a6fe4f8411018b345b03
model-00006-of-00014.safetensors	4.88 GB	1cf45c73223369e4d96908b125abb83b1592f15df7c4cc6313bc1b7634ad7ed1
model-00007-of-00014.safetensors	4.95 GB	32572fcfe8bab938df2ff753e82469164bee82b864345acfa6e01c179b5fd207
model-00008-of-00014.safetensors	4.50 GB	fe83f3b7e3128dc57545721487a09db8809ea242a5885c997fba91da3e14db7f
model-00009-of-00014.safetensors	4.59 GB	dae352f60a130eae62ab23fd2d61dc5fd1a66900716058296b9a46f003e0de03
model-00010-of-00014.safetensors	4.88 GB	70c03b933835e2acb380b60e96beaec2a76cc8d3b2f781ef61e7e7c5df31cc9e
model-00011-of-00014.safetensors	4.88 GB	0f3a552fb47e371233e29369b1a73b7db18a11dd6d598e914737e30de25808e5
model-00012-of-00014.safetensors	4.54 GB	c36fead651ecc9ad1ee1c711a1a6deaa264839cfddd346ba5912821bb55a81dd
model-00013-of-00014.safetensors	4.96 GB	59e891b65a28a8a0ebf894e80ee7b5e4e89db1ae48a690eccdd42a270e8e7c5c
model-00014-of-00014.safetensors	2.17 GB	4f618cf00485b326dec9f928b25668ea42a6b3ce95fe0ec723da3adbc4a72d1d
model.safetensors.index.json	49.2 KB	0908eeac8a5c44e06f29cd92023c1fee1c46057acb4a20edc5226950c335c184
special_tokens_map.json	601 B	35a6dd9ff6e2474ba72b180db853851d2a84fce3d12095089285f447c7474541
tokenizer.json	20.0 MB	76ebeac0d8bd7879ead7b43c16b44981f277e47225de2bd7de9ae1a6cc664a8c
tokenizer_config.json	3.19 KB	84716e4c942bf2e6de5c8834428c761ab6f3f4306fe992c1bc7791b21bbeff4a

VersionsEach time the pinned upstream commit or file list changed in this repo. Newest first.

Upstream commit	Date	Contents
af83a738fc pinned now	2026-11-15 04:15 UTC	22 files · 65.1 GB

Use it

$ fw get arcee-ai/GLM-4-32B-Base-32K
# manifest sha256 ef44ca2be1a3… · 65.1 GB · 22 files
$ fw verify ./GLM-4-32B-Base-32K arcee-ai/GLM-4-32B-Base-32K

fw.pyChecks the signature and every sha256, tries each source in order, resumes.

$ python3 fw.py get https://freeweights.allternit.com/manifests/arcee-ai/GLM-4-32B-Base-32K.json --dir ./GLM-4-32B-Base-32K
$ python3 fw.py verify ./GLM-4-32B-Base-32K ./arcee-ai__GLM-4-32B-Base-32K.json

hfThe Hugging Face CLI with one setting. Same repo id.

$ HF_ENDPOINT=https://freeweights.allternit.com hf download arcee-ai/GLM-4-32B-Base-32K --revision af83a738fc6a14b6327ddf12290c8694f807e777 --local-dir ./GLM-4-32B-Base-32K

transformersLoad the weights in Python.

import os
os.environ["HF_ENDPOINT"] = "https://freeweights.allternit.com"
from transformers import AutoModelForCausalLM, AutoTokenizer
tok = AutoTokenizer.from_pretrained("arcee-ai/GLM-4-32B-Base-32K", revision="af83a738fc6a14b6327ddf12290c8694f807e777")
model = AutoModelForCausalLM.from_pretrained("arcee-ai/GLM-4-32B-Base-32K", revision="af83a738fc6a14b6327ddf12290c8694f807e777", device_map="auto")

vLLMServe an OpenAI-compatible API.

$ HF_ENDPOINT=https://freeweights.allternit.com vllm serve arcee-ai/GLM-4-32B-Base-32K --revision af83a738fc6a14b6327ddf12290c8694f807e777

Run at home

License

Name: MIT License
Id: MIT
Class: Permissive
Redistribution: yes
Listing: Hosted (signed)
Decided: License as declared by the publisher. No person read it. 2026-11-15
Text in: README.md
Text sha256: 4df3520370d0012a06435b56425b818b208fc9fb095bb9242ed247e055ff1d01
Terms gate: No extra terms beyond the license.

Read the license text

Licenses are checked against the license table. A custom license passes only after a person reads it.

Provenance

From: arcee-ai <huggingface.co/arcee-ai>Newsgroups: alt.freeweights.glmSubject: GLM-4-32B-Base-32KDate: 2026-11-15Commit: af83a73Signed: 2DC28CC7B26A283B ✓

Upstream: arcee-ai/GLM-4-32B-Base-32K
Pinned commit: af83a738fc6a14b6327ddf12290c8694f807e777
Manifest: GLM-4-32B-Base-32K.json · .minisig
Manifest sha256: ef44ca2be1a3dbe57055d67914a8898bff809984693971d6f3187ac2e3cdcb87
Signature: minisign, Ed25519, key 2DC28CC7B26A283B
Created: 2026-11-15 04:15 UTC
Generator: freeweights-manifest/0.1.0
Takedowns: None.

How to verify

$ curl -O https://freeweights.allternit.com/manifests/arcee-ai/GLM-4-32B-Base-32K.json
$ curl -O https://freeweights.allternit.com/manifests/arcee-ai/GLM-4-32B-Base-32K.json.minisig
$ minisign -Vm GLM-4-32B-Base-32K.json -P RWQ7KGqyx4zCLXtUFOqvfv8L3g+8fpKJ1FsuxPyOYMF+4KufbUxKq1Oq
$ shasum -a 256 <file>   # compare with the Files tab

A manifest whose signature fails must be rejected. A file whose size or sha256 does not match must be discarded. Verify a download

Model tree

Parent

zai-org/GLM-4-32B-Base-0414 (not in this catalog) · this model is a finetune of it

1 model in the catalog name this one as their parent.

Quantizations: 1

Quantizations 1

bartowski/arcee-ai_GLM-4-32B-Base-32K-GGUF530 GB

Get it

Download

$ HF_ENDPOINT=https://freeweights.allternit.com hf download arcee-ai/GLM-4-32B-Base-32K

Open in Freeweights Studio

Pull, verify and run it in the Allternit app on Desktop, web or phone.

Signed manifest

minisign, key 2DC28CC7B26A283B. GLM-4-32B-Base-32K.json. Provenance. Takedown history: none.

Source

huggingface.co/arcee-ai/GLM-4-32B-Base-32K

Variants

1 quantized or adapter build of GLM-4-32B-Base-32K. Each has its own license check and page.

SubjectSizeParamsLicenseDate

arcee-ai_GLM-4-32B-Base-32K-GGUFHosted (signed) From: bartowski in alt.freeweights.glm530 GB—Permissive2026-11-15

Related models

All of alt.freeweights.glm →

From: bartowski <huggingface.co/bartowski>Subject: arcee-ai_GLM-4-32B-Base-32K-GGUF

From:

ggml-org <huggingface.co/ggml-org>Subject: AutoGLM-Phone-9B-GGUF

From:

zai-org <huggingface.co/zai-org>Subject: cogagent-chat-hf

From:

zai-org <huggingface.co/zai-org>Subject: CogView3-Plus-3B

From:

zai-org <huggingface.co/zai-org>Subject: cogvlm-base-490-hf

From: laion <huggingface.co/laion>Subject: dev_set_part1_10k_glm_4_7_traces_jupiter

GLM-4-32B-Base-32K

Works with

Runs with

Fine-tune with

Convert with

Evaluate with

Scan with

Works with

Overview

Model card

GLM-4-32B-Base-32K#

Model Details#

Improvements#

Benchmarks#

NIAH Benchmark Results Comparison#

NIAH Averages#

Use Cases#

License#

Files

Use it

License

Provenance

How to verify

Model tree

Variants

Related models