// Open-model VRAM table · family
GLM / CogVideoX (Z.ai): local deployment & VRAM
35 open-weight models from Z.ai (Zhipu) (55 official repos incl. quantised builds). At 4-bit they range from 1.9 GB (GLM-OCR) to 504 GB (GLM-5.1).
official orgs: zai-orglatest 2026-08-25memory = estimate
01Text LLM (19)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| GLM-5.3 | 753.3B | 1M | 1.66 TB | 504 GB | 8 × H100 (640GB) | — | Custom licence – read it | 2026-08-25 |
| GLM-5.2 | 753.3B | 1M | 1.66 TB | 504 GB | 8 × H100 (640GB) | — | Commercial OK | 2026-06-16 |
| GLM-5.1 | 753.9B | 198K | 1.66 TB | 504 GB | 8 × H100 (640GB) | — | Commercial OK | 2026-04-03 |
| GLM-5 | 744B / 40B | 198K | 1.64 TB | 497 GB | 8 × H100 (640GB) | — | Commercial OK | 2026-02-11 |
| GLM-4.7-Flash | 31.2B / 3B | 198K | 69.6 GB | 21.8 GB | RTX 5090 32GB | Mac 64GB | Commercial OK | 2026-01-19 |
| GLM-4.7 | 358.3B | 198K | 792 GB | 243 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB | Commercial OK | 2025-12-22 |
| GLM-4.6 | 356.8B | 198K | 789 GB | 242 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB | Commercial OK | 2025-09-29 |
| GLM-4.5 | 355B / 32B | 128K | 785 GB | 240 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB | Commercial OK | 2025-07-20 |
| GLM-4.5-Air | 106B / 12B | 128K | 235 GB | 72.7 GB | RTX PRO 6000 96GB | Mac 128GB | Commercial OK | 2025-07-20 |
| GLM-Z1-9B-0414 ↗ | 9.4B | 32K | 21.5 GB | 7.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-04-08 |
| GLM-Z1-32B-0414 | 32.6B | 32K | 72.7 GB | 22.7 GB | RTX 5090 32GB | Mac 64GB | Commercial OK | 2025-04-08 |
| GLM-4-9B-0414 | 9.4B | 32K | 21.5 GB | 7.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-04-07 |
| GLM-4-32B-0414 | 32.6B | 32K | 72.7 GB | 22.7 GB | RTX 5090 32GB | Mac 64GB | Commercial OK | 2025-04-07 |
| glm-edge-1.5b-chat | 1.59B | 8K | 4.5 GB | 2.0 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2024-11-20 |
| glm-edge-4b-chat | 4.33B | 8K | 11.0 GB | 4.4 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2024-11-20 |
| codegeex4-all-9b | 9.4B | 128K | 26.6 GB | 12.1 GB | 16GB GPU | Mac 24GB | Custom licence – read it | 2024-07-05 |
| glm-4-9b-chat | 9.4B | 128K | 26.6 GB | 12.1 GB | 16GB GPU | Mac 24GB | Custom licence – read it | 2024-06-04 |
| glm-4-9b-chat-1m | 9.48B | 1M | 26.7 GB | 12.2 GB | 16GB GPU | Mac 24GB | Custom licence – read it | 2024-06-04 |
| glm-4v-9b | 13.9B | 8K | 36.5 GB | 15.1 GB | RTX 4090 24GB | Mac 36GB | Custom licence – read it | 2024-06-04 |
02Vision-language (8)
| Model | Params | Context | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|---|
| GLM-5.3-Flash | 320B / 18B | 1M | 705 GB | 214 GB | 4 × RTX PRO 6000 (384GB) | Mac 512GB | Commercial OK | 2026-08-25 |
| GLM-OCR | 1.32B | 128K | 4.0 GB | 1.9 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-01-30 |
| AutoGLM-Phone-9B | 9B | 64K | 20.6 GB | 6.8 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-12-08 |
| GLM-4.6V | 107.7B | 128K | 239 GB | 73.9 GB | RTX PRO 6000 96GB | Mac 128GB | Commercial OK | 2025-12-07 |
| GLM-4.6V-Flash | 10.3B | 128K | 23.5 GB | 7.7 GB | 12GB GPU | Mac 16GB | Commercial OK | 2025-12-07 |
| GLM-4.5V | 107.7B | 64K | 239 GB | 73.9 GB | RTX PRO 6000 96GB | Mac 128GB | Commercial OK | 2025-08-10 |
| GLM-4.1V-9B-Thinking | 10.3B | 64K | 23.5 GB | 7.7 GB | 12GB GPU | Mac 16GB | Commercial OK | 2025-06-28 |
| cogvlm2-llama3-chat-19B | 19.5B | 8K | 47.7 GB | 17.8 GB | RTX 4090 24GB | Mac 36GB | Custom licence – read it | 2024-05-16 |
03Speech-to-text (1)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| GLM-ASR-Nano-2512 | 2.26B | 5.5 GB | 2.0 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-12-09 |
04Image generation (2)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| GLM-Image | 6.93B | 15.7 GB | 5.1 GB | 8GB GPU | Mac 16GB | Commercial OK | 2026-01-08 |
| CogView4-6B ↗ | 6.37B | 14.5 GB | 4.7 GB | 8GB GPU | Mac 16GB | Commercial OK | 2025-03-03 |
05Video generation (5)
| Model | Params | BF16 | 4-bit | Smallest GPU (4-bit) | Mac (4-bit) | Licence | Released |
|---|---|---|---|---|---|---|---|
| SCAIL-2 | — | — | — | — | — | Commercial OK | 2026-06-09 |
| CogVideoX1.5-5B-I2V ↗ | 5.57B | 12.8 GB | 4.2 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2024-11-02 |
| CogVideoX-5b-I2V | 5.63B | 12.9 GB | 4.3 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2024-09-16 |
| CogVideoX-5b | 5.57B | 12.8 GB | 4.2 GB | 8GB GPU | Mac 16GB | Custom licence – read it | 2024-08-17 |
| CogVideoX-2b | 1.69B | 4.2 GB | 1.6 GB | 8GB GPU | Mac 16GB | Commercial OK | 2024-08-05 |
06Other families
Qwen · IBM Granite · NVIDIA Nemotron · Gemma · Mistral / Magistral / Devstral · DeepSeek · Ai2 OLMo / Molmo · Ant Ling / Ring / Ming · Liquid LFM · MiniCPM / VoxCPM · Tencent Hunyuan / Hy · Phi · Llama · Meituan LongCat · Jina embeddings & rerankers · Wan · Kimi · Cohere Command / Aya · Xiaomi MiMo · FLUX · Microsoft VibeVoice / Fara / Harrier · InternLM / Intern-S · StepFun Step · MiniMax · Baidu ERNIE · NVIDIA Parakeet / Canary speech · Stable Diffusion / Stable Audio · Yi (01.AI) · ByteDance Seed · SmolLM / SmolVLM · BGE embeddings & rerankers · LTX Video · FunAudioLLM speech · gpt-oss · poolside Laguna · Whisper · Baichuan · Thinking Machines Inkling · xAI Grok · rednote dots
Totals include 10% + 0.5GB overhead and one 8K-token KV cache for text models. Estimates only. Method →