Glm 4 9B Chat — Hardware Requirements & GPU Compatibility
ChatFunctions
GLM-4-9B-Chat is Zhipu AI's 9.4-billion-parameter chat model, serving as the human-preference-aligned iteration of the open GLM-4-9B and belonging to the fourth-generation GLM series. Beyond handling multi-turn conversations, it offers web browsing, code execution, custom tool calling, and long-text reasoning, supporting 26 languages like Japanese, Korean, and German. The card indicates it outperforms Llama-3-8B-Instruct on benchmarks including MMLU, MATH, and HumanEval. At this scale, it runs on a single consumer GPU after quantization. Its context window spans 131,072 tokens, aligning with the card's 128K claim, while a separate variant supports up to 1M tokens. Released under the custom GLM-4 license instead of a standard open-source one, its terms warrant review before commercial use. Published in June 2024, the repository employs custom ChatGLM code, and the card recommends using the glm-4-9b-chat-hf version for newer transformers releases.
Runs great on a 24 GB GPU
Runs from 8 GB (Q4_K_M · 8K)Best on 24 GB (Q4_K_M · ~106 tok/s · 64K)Macs from 16 GB (Q4_K_M · ~23 tok/s)
Other
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Glm 4 9B Chat
One recommended way. Every other method, OS and quant is in the step-by-step guide.Each way opens a step-by-step guide for your OS and quant.
Recommended
ollama run glm4:9b-chat-q4_K_MQ4_K_M fits fully on the RTX 4090 · 15.1 GB free · 16K context
Download: glm-4-9b-chat-Q4_K_M.gguf6.25 GBbartowski
Ways to runOther ways
each opens step-by-stepRun in cloud
Runs on RTX 3080 Ti (12 GB) (3 GB headroom) · Q4_K_M
- Generation speed
- ~96 tok/s
- generation speed
- Cost per 1M output tokens
- $0.35
- per 1M output tokens
How Much VRAM Does Glm 4 9B Chat Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 4.4 GB | 10.1 GB | 2-bit · K-quant — 2-bit quantization with K-quant improvements | Fits: 64K context · ~149 tok/s | |
| 4.5 GB | 10.3 GB | 3-bit small — 3-bit small quantization | Fits: 64K context · ~145 tok/s | |
| 5.0 GB | 10.8 GB | 3-bit medium — 3-bit medium quantization | Fits: 64K context · ~130 tok/s | |
| 5.2 GB | 10.9 GB | 4-bit · legacy — 4-bit legacy quantization | Fits: 64K context · ~127 tok/s | |
| 6.2 GB | 12 GB | 4-bit medium — 4-bit medium quantization — most popular sweet spot | Fits: 64K context · ~106 tok/s | |
| 7.4 GB | 13.1 GB | 5-bit medium — 5-bit medium quantization — good quality/size tradeoff | Fits: 64K context · ~89 tok/s | |
| 8.5 GB | 14.3 GB | 6-bit, very good — 6-bit quantization, very good quality | Fits: 64K context · ~77 tok/s | |
| 10.3 GB | 16.1 GB | 8-bit, near-lossless — 8-bit quantization, near-lossless | Fits: 64K context · ~63 tok/s |
Selected Q4_K_M
4-bit medium quantization8.9GBat 16K context
- Weights
- 5.5 GB
- KV cache
- 3.1 GB
- Overhead
- 0.3 GB
Fits RTX 4090 (reference) · 15.1 GB free · ~106 tok/s
≥50 tok/s needs ≥476 GB/s · RTX 5090 ~188, RTX 3090 Ti ~106
Which Hardware Can Run Glm 4 9B Chat?
Q4_K_M · 6.2 GBGlm 4 9B Chat (Q4_K_M) requires 6.2 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 9+ GB is recommended. 52 GPUs can run it, including NVIDIA GeForce RTX 5090, NVIDIA GeForce RTX 3090 Ti, NVIDIA GeForce RTX 3070 Ti. 151 devices with unified memory can run Glm 4 9B Chat, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), NVIDIA Jetson Orin Nano 8GB (Super).
12 GB
12 GB tier →fits · 3.1 GB free · up to 32K ctx
10 GPUs · 1 device family
RTX 3080 Ti~96Simulate output on RTX 3080 TiRTX 5070~71Simulate output on RTX 5070RTX 4070~53Simulate output on RTX 4070iPhone~9Simulate output on Apple iPhone 17 Pro20 GB
fits · 11.1 GB free · up to 64K ctx
1 GPU
RX 7900 XT~77Simulate output on RX 7900 XT32–40 GB
fits · 23.1 GB free · up to 128K ctx
1 GPU · 3 Mac classes · 4 device families
RTX 5090~188Simulate output on RTX 5090M5 Max36 GB~52Simulate output on Mac Studio M5 Max (36 GB)M532 GB~17Simulate output on MacBook Air 13" M5 (32 GB)Orin~22Simulate output on NVIDIA Jetson AGX Orin 32GBSnapdragon X Elite (X1E-80-100)~14Simulate output on Snapdragon X Elite Copilot+ PC
By VRAM tier · typical GPUs and Macs at Q4_K_M
How good is Glm 4 9B Chat?
llmrun Score blends public benchmark boards into one number; 100 = the average open model.
Not enough data yet for an llmrun Score.
A Score needs results on enough independent boards. How the Composite Score is built
Benchmarks
This is the single board showing results for this model; badges show who ran the test.
- MMLU-Pro
TIGER-LabIndependent
Independent48.0%
#78/146 open
Alternatives to Glm 4 9B Chat
GLM 4 family · 18 models →Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| SenseNova Loop 9B Previewsensenova | — | 8.2 GB | ~103 | same size |
| Granite Switch 4.1 8BIBM | — | 8.4 GB | ~111 | same size |
| ZDTaichu5.0 9BTaichuAI | — | 9.2 GB | ~101 | same size |
| Ornith 1.5 9Bornith-ai | — | 6.6 GB | ~103 | same size |
| Apertus V1.5 8Bswiss-ai | — | 8.5 GB | ~111 | same size |
| ZAYA1 VL 8BZyphra | — | 7.5 GB | ~346 | same size |
Specifications
- Publisher
- Z.ai
- Family
- GLM 4
- Parameters
- 9.4B
- Architecture
- ChatGLMModel
- Context Length
- 131,072 tokens
- Release Date
- 2024-06-04
- License
- Other
Frequently Asked Questions
How much VRAM does Glm 4 9B Chat need?
Glm 4 9B Chat requires 6.2 GB of VRAM at Q4_K_M, or 20.7 GB at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 9.4B × 4.8 bits ÷ 8 = 5.6 GB
KV Cache + Overhead ≈ 0.6 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
VRAM usage by quantization
What's the best quantization for Glm 4 9B Chat?
For Glm 4 9B Chat, Q4_K_M (6.2 GB) offers the best balance of quality and VRAM usage. Q4_K_L (6.3 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XXS at 2.8 GB.
VRAM requirement by quantization
★ Recommended — best balance of quality and VRAM usage.
Can I run Glm 4 9B Chat on a Mac?
Yes — Mac Mini M2 Pro (16 GB) and 133 other Macs can run Glm 4 9B Chat. Apple Silicon uses unified memory, so the model shares RAM with the system. At Q4_K_M you need at least 6.2 GB of usable unified memory (RAM minus macOS overhead).
Can I run Glm 4 9B Chat locally?
Yes — Glm 4 9B Chat can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Glm 4 9B Chat?
At Q4_K_M, Glm 4 9B Chat can reach ~774 tok/s on AMD Instinct MI350X. On NVIDIA GeForce RTX 4090: ~106 tok/s. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 6.2 × 0.65 = ~839 tok/s
Estimated speed at Q4_K_M (6.2 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Glm 4 9B Chat?
At Q4_K_M, the download is about 5.64 GB. The full-precision BF16 version is 18.80 GB. The smallest option (IQ2_XXS) is 2.58 GB.
Which GPUs can run Glm 4 9B Chat?
52 consumer GPUs can run Glm 4 9B Chat at Q4_K_M. Top options include AMD Radeon RX 6700 XT, AMD Radeon RX 6800, AMD Radeon RX 6800 XT, AMD Radeon RX 7600. 40 GPUs have plenty of headroom for comfortable inference.
Which devices can run Glm 4 9B Chat?
151 devices with unified memory can run Glm 4 9B Chat at Q4_K_M (6.2 GB), including AMD Ryzen AI 9 HX 370 (Strix Point) Laptop, ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.