Smaller alternatives
Models that need less memory than GLM 4.7 Flash and score about as well or better.
GLM 4.7 Flash
MoE
Z.ai · 31.2B (3B active) · Mixture of experts
GLM 4.7 Flash is a 31.2B model from Z.ai with a 198K-token context. At Q4_K_M with an 8K context it needs about 17.8 GB; 24 GB leaves room for longer chats.
Mixture of experts
Parameters: 31.2B · Active: 3B
All 31.2B parameters load into memory, but only 3B work on each token, so it runs at the speed of a much smaller model.
Quantization options
| Quant | Bits | Memory | Quality | On your hardware |
|---|---|---|---|---|
| Q2_K | 3.16 | — | Noticeably worse | … |
| Q3_K_M | 4 | — | Some loss | … |
| Q4_K_M | 4.89 | — | Good | … |
| Q5_K_M | 5.7 | — | Good | … |
| Q6_K | 6.56 | — | Near original | … |
| Q8_0 | 8.5 | — | Near original | … |
| F16 | 16 | — | Original | … |
Memory by context length
At Q4_K_M. The context cache grows with every token the model keeps in mind.
| Context | Context cache | Total |
|---|---|---|
| 4K | 0.2 GB | 17.6 GB |
| 8K | 0.4 GB | 17.8 GB |
| 32K | 1.7 GB | 19 GB |
| 128K | 6.6 GB | 24 GB |
Next steps
Compare it with
Scored models of a similar size, side by side on your device.
Least hardware that runs it well
Desktop devices with the least memory that give GLM 4.7 Flash grade A or S at Q4_K_M.
Can I run GLM 4.7 Flash locally?
Can I run GLM 4.7 Flash locally?
Yes, if your GPU or Mac has about 17.8 GB free for it at Q4_K_M. Open this page on that computer to see the grade for your exact hardware, then start it with llama.cpp, Ollama or LM Studio.
How much memory does GLM 4.7 Flash need?
About 17.8 GB at Q4_K_M with an 8K context and 30.4 GB at Q8_0. Lower quantizations fit smaller cards with some loss in quality; longer contexts add to the total.