// GLM / CogVideoX (Z.ai) · Text LLM
GLM-4.7-Flash: VRAM & local deployment
31.2B total · 3B active (MoE) parameters. About 21.8 GB at 4-bit and 69.6 GB at BF16 with one 8K-token request.
released 2026-01-19licence: Commercial OKmemory figures are estimates
01Specs
| Publisher | Z.ai (Zhipu) · zai-org |
|---|---|
| Family / series | GLM / CogVideoX (Z.ai) · GLM-4.7-Flash |
| Type | Text LLM (text-generation) |
| Parameters | 31.2B total · 3B active (MoE) Source: counted from official safetensors weightsActive parameters from the model card: “GLM-4.7-Flash is a 30B-A3B MoE model.” |
| Context length | 198K tokens (202,752)max_position_embeddings in config.json |
| Architecture | 47 layers · MLA · MoE |
| Released weights | BF16 · 62.4 GBdownload |
| Licence | MIT · Commercial OKPermissive licence (Apache-2.0 / MIT / CC-BY / OpenMDW): commercial use allowed. |
| Released | 2026-01-19 (Hugging Face repo created) |
02How much VRAM it needs (estimate)
Total = weights + 10% + 0.5GB overhead + KV cache for one request (FP16 KV). More concurrent users multiply the KV column.
| Precision | Weights | Total @ 8K | Total @ 32K | Total @ 128K | Smallest GPU | Mac |
|---|---|---|---|---|---|---|
| BF16 / FP16 (original) | 62.4 GB | 69.6 GB | 71.0 GB | 76.3 GB | H100 80GB | Mac 128GB |
| FP8 / INT8 / Q8_0 | 33.2 GB | 37.4 GB | 38.8 GB | 44.1 GB | 2 × RTX 5090 (64GB) | Mac 64GB |
| 4-bit (Q4_K_M / AWQ / GPTQ-Int4) | 18.9 GB | 21.8 GB | 23.1 GB | 28.4 GB | RTX 5090 32GB | Mac 64GB |
KV cache per request: 8K 0.4 GB · 32K 1.8 GB · 128K 7.1 GB
Bits per weight: BF16 16, FP8 8.5, 4-bit 4.85. “Fits” = total ≤ 90% of memory; Mac GPU can use ~2/3 of unified memory (≤36GB) or ~3/4 (larger). For MoE, all experts must be in memory, so size by total parameters. Method →
03How to run it
Only runtimes the publisher or the runtime’s official library names for this model.
- vLLM — mentioned in the official model card
- SGLang — mentioned in the official model card
- Transformers — mentioned in the official model card
- Ollama — in the official Ollama library · glm-4.7-flash
ollama run glm-4.7-flash04Official repositories & quantised builds
1 official repo(s) from zai-org.
- GLM-4.7-Flash main
05Other GLM / CogVideoX (Z.ai) models
All 35 →06Sources
Data checked 2026-01-29. Parameters, context and licence are copied from the official repository; memory figures are our estimates.